(dissenting).
That bis (2,6-dimethyImorphoIino-thiocarbonyl) monosulfide, appellants’ vulcanization accelerator, is not specifically disclosed by the references cannot be disputed. That the references, taken in combination, do fairly teach that *382thiuram sulfides1 (monosulfides, disul-fides, and polysulfides) are in general good vulcanization accelerators for GR-S (butadiene-styrene copolymer) and can be used in GR-S with less danger of “scorching”2 than in natural rubber, also cannot be disputed.
However, it is urged by appellants that they have demonstrated that their particular vulcanization accelerator has an unexpectedly high degree of processing safety. Specifically, appellants compared the processing safety of their accelerator to that of four other closely related thiuram sulfides, including the most closely related compound specifically disclosed in Naunton. Five compounded butadiene-styrene copolymer3 compositions were prepared. The compositions were identical except for the thiuram sulfide included. “Mooney scorch data” were obtained in terms of “minutes to 10-point rise.” Appellants have indicated that this is “a standard test in determining the ‘safety’ of a composition using a particular curing agent,” and this has not been disputed.
The values obtained are shown in Table I.
Table I
Thiuram Sulfide Minutes to 10-Point Rise
R-CS-S-CS-R* where R is:
1. 2,6-dimethylmorpholino 36
2. morpholino 20
3. 2,5-dimethylmorpholino 22
4. 3,5-dimethylmorpholino 18
R-CS-Sz-CS-R** where R is:
5. 2,6-dimethylmorpholino 14
* thiuram monosulfides
** thiuram disulfides
To compare the rapidity and quality of cure obtained with appellants’ accelerator and that obtained with compounds 2, 3 and 4 above, compounded butadiene-*383styrene compositions prepared in the same manner were “press-cured” and the tensile properties of each sample were measured by an ASTM method. Table II shows the data obtained in this test,
Appellants’ specification asserts that the above data demonstrates the following:
From Table II it can be seen that the 2,6-dimethyl and 3,5-dimethyl isomers and the morpholino homolog are relatively equivalent in rapidity and quality of cure but that the 2,5-di-methyl isomer gives a slow and unsatisfactory cure. However, as shown in Table I, only the 2,6-dimethyl isomer combines fast action and satisfactory cure with an exceptionally high degree of processing safety.
Note should be taken that the morpholino thiuram monosulfide (compound 2) is the only thiuram monosulfide specifically disclosed in Naunton, that compounds 3 and 4 are position isomers of appellants’ accelerator, and that compound 5 is the 2,6-dimethyl analog of the only thiuram-cfeulfide specifically disclosed in Naun-ton (morpholine thiuram disulfide). It will be observed also that, compared to compounds 2 through 5, appellants’ accelerator provides 80%, 60%, 100% and 150% higher processing safety, respectively.
The examiner, in his Answer, treated appellants’ showings as follows:
While the Examiner does not disagree with the Mooney scorch data comparisons, per se, the relative processing safety of the monosulfides and the disulfide would be expected and any differences — one of degree only. It is evident from the applied references, of the obvious use of the thiuram sulfide compounds and it would be a matter of routine experimentation to determine the best accelerator for one’s use. Since it is quite clear that the claimed bis (2,6-dimethylmor-pholinothiocarbonyl) monosulfide would be expected to effect the desired results in a GR-S composition, as taught by Naunton et al. in view of Payman et al. and Whitby, merely testing to determine the relative degrees of processing safety and finding a particular thiuram sulfide exhibiting *384a higher or lower degree would be expected. It is well settled that obviousness does not require [absolute] predictability; In re Moreton 1961 C.D. 330 [48 CCPA 928, 288 F.2d 940, 129 USPQ 288]. [Emphasis added.]
First, in In re Krazinski, 347 F.2d 656, 52 CCPA 1447 (1965), this court, in supporting a unanimous decision, said:
* * * the rejection is based upon section 103, which says nothing at all about differences in degree and differences in kind, but instead clearly requires us, in determining whether a patentable invention has been made, to consider the differences, whatever their nature, between the subject matter sought to be patented and the prior art and to determine if the subject matter as a whole would have been obvious to one of ordinary skill in the art at the time the invention was made. [Emphasis quoted.]
Likewise, § 103 says nothing about “routine experimentation” or “mere testing”; it does, however, say something that apparently needs repeating, namely :
Patentability shall not be negatived by the manner in which the invention was made. [Emphasis added.]
Second, Moreton does not, on its facts, support the rejection here. There the art taught that polymerized methacrylie acid esters in which the ester portion contains from 2 to 15 carbon atoms are useful as “viscosity improving agents” in hydraulic fluids; applicant was claiming polymerized decyl methacrylate (10 carbon atoms in the ester portion) in combination with a known hydraulic fluid. Moreton did not purport to show that the polymerized methacrylie acid ester he used was, for his purpose, unexpectedly superior to any of the other polymerized methacrylie acid esters which the art generically disclosed. Instead, he merely argued that certain prior art, not relied on by the examiner, demonstrated that combinations of known viscosity improving agents and known lubricants and hydraulic fluids had sometimes performed unsatisfactorily. It was in this context that we said in Moreton, 288 F.2d at 943, at 48 CCPA 933:
Obviousness does not require absolute predictability. Where, as here, the knowledge of the art clearly suggests a certain class of compounds, materials actually known by the term “viscosity improving agents,” as useful to improve the viscosity index of a certain group of hydraulic fluid lubricants, the mere possibility of failure does not render their successful use “unobvious.”
With respect to appellants’ comparative tests, the board had this to say:
* * * [they] provide uncertain evidence of the superiority of appellants’ accelerator as compared with even the arbitrarily selected homologs of the prior art.- The experiments appear to deal with a single butadiene-styrene copolymer, one having 23.5% bound styrene, and have not been shown to have a general significance as applied to the butadiene-styrene copolymers as a group. [Emphasis added.]
Although the board attacks appellants’ choice of thiuram sulfides for comparative tests, no hint is given as to why the choice was thought to be improper or what compounds the board would have had appellants employ. The board appears to have overlooked the fact that appellants have compared their accelerator with the only thiuram monosulfide specifically disclosed in the cited art and with a thiuram disulfide more closely related to their own accelerator than the only disulfide specifically disclosed in the cited art. It should also be noted that there are only four non-equivalent ways in which two methyl groups can be introduced onto the four carbon atoms of morpholine, viz., 2,3-, 2,5-, 2,6-, and 3.5- (3,6- and 5,6- being the same as 2.5- and 2,3-, respectively), and that appellants have established that their isomer is superior to two of the remaining three isomers. It would seem, then, that having taken one step in each di*385rection from the accelerator employed in their composition, appellants have done all that can fairly be asked of them in selecting accelerator compounds for comparison.
With respect to the board’s second criticism of appellants’ experiments, namely, the specific rubber selected for testing, I find no basis of record for speculating that the superiority which has been shown as to the copolymer used would not have general significance to butadiene-styrene copolymers as a group. The examiner apparently saw nothing wrong with this aspect of the experiments or, for that matter, appellants’ choice of compounds for comparison; in fact, he said, “the Examiner does not disagree with the * * * comparisons, per se * * (Emphasis added.)
The majority opinion dismisses appellants’ showings stating:
We agree with the board and solicitor that the comparative tests, limited to a particular copolymer of butadiene-styrene, are insufficient to establish similar superiority for the accelerator with all copolymers of butadiene-styrene within the scope of the claim and are inadequate to overcome the strong suggestion of obviousness inferred from the references.
Moreover, we agree with the examiner that, inasmuch as the art teaches that the morpholine thiuram mono-sulfides and disulfides are “safe” accelerators for vulcanizing rubber and since it would not be unreasonable to expect that individual accelerators would tend to vary in this effect with the same or different rubbers, “merely testing to determine the relative determine the relative degrees of processing safety and finding a particular thiuram sulfide exhibiting a higher or lower degree would be expected.” Obviousness does not require absolute predictability. * * * We are unable to find “clear and convincing evidence,” * * * that appellants’ accelerator possesses unexpected processing safety as compared to the other thiuram monosulfides. [Emphasis added.]
Boiled down to its bare bones, the majority’s position side-steps a consideration of appellants’ invention as a whole by employing, as did the examiner, the “mere testing” cliché, thereby avoiding even an acknowledgment of the magnitude of the superiority which appellants have shown. While minor variations in a given property are to be expected as one goes through a homologous series or compares compounds very closely related structurally, for example the type variations shown by the data for compounds 2, 3, and 4 in Table I above, I cannot agree that the 60% to 150% differences which appellants have shown are either minor or suggested by the art. Certainly they are as unobvious as they are unpredictable.
The observation of the board and the majority that only one copolymer was used in appellants’ experiments appears to be nothing more than that, an observation. The examiner did not urge that the showings were defective in this respect, and, as stated above, there is no basis in the record for speculating that, if appellants’ experiments were repeated using a butadiene-styrene copolymer containing a different percentage of bound styrene, the superiority of appellants’ accelerator would be demonstrated any less dramatically.
Thus, although the examiner established a prima facie case of obviousness by citing Naunton, Payman, and Whitby, appellants have satisfactorily rebutted that case by demonstrating that their vulcanization accelerator possesses a degree of processing safety so significantly superior to that possessed by the other closely related compounds specifically and/or generically disclosed by the prior art that that superiority exceeds that which might reasonably be expected by persons of ordinary skill in this art. Appellants have, therefore, made a patentable advance in the art. Considering appellants’ invention as a whole, I find it to be non-obvious and would reverse.
. That is, compounds containing the grouping
where “x” is an integer of one or more.
. This term is explained in the following paragraph from appellants’ specification:
Among the technically valuable attributes of an accelerator for the vulcanization of natural and synthetic rubber are (a) delayed action and (b) the capacity to activate a rapid, efficient, and stable curing reaction leading to a high quality vulcanízate. The former attribute is essential to the avoidance of premature vulcanization or scorching of rubber stocks during compounding, shaping, and molding operations. The latter is necessary in order to minimize the time required for the curing step of the overall manufacturing process and to produce the highest possible quality in the finished product. [Emphasis added.]
. The copolymer employed was identified as “SBR-1500 (previously known as GR-S-1500)” which is said to contain “23.5% bound styrene.” Appellants’ specification states that “The most commonly used butadiene-styrene copolymers contain about 76.5 percent of butadiene and about 23.5 percent styrene, although other butadiene-styrene copolymers may contain up to about 45 percent styrene.”