delivered the opinion of the court:
This appeal is from the decision of the Patent Office Board of Appeals affirming the rejection of claim 1, the only remaining claim of appellants’ application No. 291,822, for a patent entitled “Method of Manufacture of Opaque Glass in Ribbon Form” on the ground of lack of invention over prior art.
The appealed claim reads:
1. A continuous method of manufacturing opaque glass in ribbon form from a slow setting glass composition which comprises continuously drawing molten glass from a reservoir, continuously casting it in ribbon form and continuously feeding the formed ribbon towards a supporting surface, and forming by impression on the reverse face of the ribbon while in plastic form and while in transit towards said supporting surface intersecting sets of reinforcing ribs of thread like nature, the height of the ribs from the reverse face being about .03" to .06",-the faces of the ribs being mutually divergent at an angle which may vary from about 135° to about 90°, each set of ribs being spaced at a pitch of about seven to the inch, the ratio of a total surface of the reverse face to the area of the obverse face being approximately between 1.1 to 1 and 1.4 to 1 whereby an accelerated cooling and setting of the glass in the ribs is achieved, and an integral stiffening mesh of chilled glass is produced on the reverse face during the continuous movement of the ribbon towards said supporting surface.
*1019The references relied on are:
Walsh (Design), 44,948, Nov. 25,1913.
Ferris, 83,841, Nov .10,1868.
Brogan et al., 370,176, Sept. 20,1887.
Nameche, 1,928,026, Sept. 26,1933.
Appellants’ application discloses a process of the type in which molten glass passes, from a tank between spaced upper and lower rolls which form it into a sheet of the desired thickness, after which it passes over a short inclined surface and onto a horizontal cooling support. Upon leaving that support it is transported by rollers to a lehr where it is annealed. The process is continuous and the glass is in the form of a uniformly moving sheet, rather than in batches.
It is stated in appellants’ specification that previous attempts to manufacture slow setting glass by a process of the kind above described, which is an old process as applied to quick setting glass, have been unsuccessful, apparently due to the fact that the glass remains so soft that it becomes distorted in passing from the forming rolls to the cooling surface,- which distortion is increased during cooling.
Appellants avoid such distortion of the glass by providing their lower roll with two intersecting sets of grooves which form on the under surface of the glass passing over the roll intersecting sets of upstanding ribs of the proportions specifically set forth in the appealed claim. The provision of these ribs not only increases the exposed surface area, with a resultant increase in the rate of cooling, but also increases the rigidity of the glass, so that distortion of the ribbon between leaving the rolls and deposit on the cooling surface is eliminated. Appellants’ specification indicates that the particular size and spacing of the ribs as set forth in the appealed claim, are essential to proper operation.
The Brogan et al. patent relates to the manufacture of checkered glass by a batch process in which a sheet of molten glass is cast upon a table having a checkered surface and is rolled with a smooth roller, so that its lower surface conforms to the surface of the table. After the glass has cooled enough to be partially stiffened it is momentarily lifted, by hand, at one end “to admit air under it and prevent dimming of the checkered surface of the glass by the rapid cooling action of the table.” The patent also suggests, as less satisfactory, an alternative prior art procedure in which the roller has a checkered surface and the table a smooth one.
The Walsh patent merely shows a design for checkered sheet glass in which there are grooves where appellants have ribs and adds nothing material to what is disclosed by Brogan et al.
The patent to Ferris discloses a continuous process of making sheet glass in which the glass is formed between rollers and passes onto a *1020horizontal cooling surface. It is stated in the patent- specification that the glass may be ornamented by engraving any desired design upon either roller “to produce either relief or countersunk designs in the glass.” -
The hi ameche patent also relates to a continuous process of making sheet glass and discloses an arrangement of glass tank, upper and lower rolls and supporting and cooling surfacés • similar to those shown by appellants. • The surfaces of the roils are smooth and no provision is made for forming ribs or other protuberances on either surface of the glass.
In considering the claim, it is to be noted that appellants’ specification defines the “reverse” face of the glass as the one which “lies on the table” or, in other words, the lower face.
The claim was rejected on Brogan et al. on the ground that it would be obvious in view of the teaching of that patent to form checkered glass in a continuous sheet by correspondingly shaping one of the rolls by which the sheet was formed, and was further rejected on Ferris or Nameche in view of Walsh on the ground that it would not require invention to provide one of the rollers of either of the former patents with the design disclosed by the latter.
While the rejections involve specifically different selections of features from the various references, they are actually based on a single proposition, which may be stated as follows: (1) Slow setting opaque glass is old; (2) It is old to form quick setting sheet glass by a continuous process in which molten glass is passed between upper and lower rolls, either of which may be engraved to produce a desired ornamental design on the glass, after which the glass is cooled on a. supporting surface; (3) It is old to ornament sheet glass with a pattern of intersecting sets of ribs; and (4) It would be obvious to select a design of intersecting ribs of the dimensions specified in the appealed claim and to apply it to the lower surface of a sheet of slow setting glass by impressing it on the lower roll. The first three statements are correct and it is only necessary to determine whether they support the fourth.
Superficially it might appear that the kind of glass, specific design and particular roll to which the design is applied are merely matters of choice; and that would doubtless be the case if the particular features selected did not combine to produce a new result not suggested by the references. Such a result, however, is clearly produced by the specific combination set forth in the appealed claim.
It is stated in appellants’ specification that previous attempts to form sheets of slow setting opaque glass by continuous processes were unsuccessful due to distortion of the glass. Neither the examiner nor the board has disputed that statement and no reference *1021showing the formation of such glass by a continuous process has. been cited. It appears, therefore, that by the claimed combination of steps, appellants have been able to solve a problem and produce a desirable result not heretofore obtainable. Under such circumstances, it is immaterial that the claimed steps may be old individually.
. [2] While it is true that patentability may not be predicated on a result which would flow naturally from the teaching of the prior art, even though the art does not suggest that particular result or its desirability, In re Kepler, 30 C. C. P. A. (Patents) 726, 132 F. 2d 130, 56 USPQ 177, the process recited by the appealed claim would not naturally be produced by following the teachings of the references. They do not even recognize the existence of a problem in the continuous production of slow setting glass.
There is nothing in any of the references to suggest that the placing-of ribs on sheet glass tends -to avoid distortion or facilitates the manufacture of the glass in any way. Accordingly, since the application of a continuous process employing rolls to smooth slow setting glass had been found not to be feasible, there would be no reason to suppose that such glass could be produced by that process if a design were impressed on it.
In holding that the application of the claimed method to slow setting glass was not inventive, the board relied on In re Ducci, 42 C. C. P. A. (Patents) 1088, 225 F. 2d 683, 107 USPQ 88, in which it was held that the application to glass having a particular linear coefficient of expansion of a process previously applied to other types of glass was unpatentable. In that case, however, the court pointed out that appellants’ glass was analogous to the glass of the references and that “it is quite evident that it can be converted to multicellular glass by the method disclosed in the references.” Here, on the other hand, it clearly could not be evident that a continuous process which had been successful with quick setting glass would also be applicable to slow setting glass, since previous attempts to apply the continuous process to the slow setting glass had not been satisfactory.
While each of the steps of appellants’ process is old, individually, their combination in the specific manner claimed is not disclosed or fairly suggested by the references and it produces a new and un-obvious result. The appealed claim is commensurate in scope with their invention and fairly defines their contribution to the art.
Certainly it cannot be said that the soundness of the rejection is clear beyond doubt; and doubts in such cases are to be resolved in favor of applicants. In re Hummer, 44 C. C. P. A. (Patents) 814, 241 F. 2d 742, 113 USPQ 66; In re Citron, 45 C. C. P. A. (Patents) 773, 251 F. 2d 619, 116 USPQ 409.
The decision of the Board of Appeals is reversed.