ALARCON, Circuit Judge:
Appellants Winn-Proof Corporation, William Werner, and Wer-Nel Enterprises, Inc. appeal from a judgment of the district court finding claims 3-5 of United States Letters Patents No. 3,646,696 [“the ’696 patent”] valid and infringed. Appellee Robert Sarkisian cross-appeals from the court’s judgment finding United States Letter Patent No. 3,662,482 [“the ’482 patent”] invalid for double patenting. For the reasons stated below, we affirm the judgment of validity and infringement but reverse the finding as to double patenting.
The invention involved in this case is a portable sign stand of the sort commonly found at roadside construction sites and gasoline stations, used to display gasoline prices and other roadside warning messages. Despite the fact that the sign stand is light and portable, with a virtually weightless base, the sign stand can withstand wind conditions of up to 80 miles per hour without “walking” or tipping over.
It is the unique interaction among the elements of the sign stand which enable it to combine the features of lightness and stability. The lightweight base of the stand is made up of tubular members or legs, arranged in either a parallel relationship or in a pyramid-like arrangement. (See Appendix, Figs. 1 and 2). The display boards of the sign stand are attached to the sign’s base by a pair of non-concentric tension springs, similar to those found on gate closing devices. The “initial” or “pre-loaded” degree of compression of the springs, measured by the amount of force required to separate their coils, is set in relation to other dimensions of the sign stand, such as the length and weight of the base, the location of the sign stand’s center of gravity, and the dimensions and weight of the display board. Under ordinary conditions, the springs are sufficiently rigid to hold the display board in an upright position. The springs’ resilience is such, however, that the force required to uncoil the springs and deflect the display board attached to them is less than the force required to topple the entire sign stand over. As a result, the sign stand will not tip over in high winds, although the springs will bend and the display board deflect. Because of the operation of the springs in relation to the other dimensions of the sign stand, the base of the sign stand can be extremely lightweight without decreasing the sign stand’s stability-1
In early 1975, defendant Werner developed the spring mounted sign stand shown in figure 3 of the Appendix. Werner’s sign stand and the three different sign stands which succeeded it, (See Appendix, Figs. 4-6) utilized the same operating principles as Sarkisian’s sign stand, though they differed from Sarkisian’s in appearance. While Werner’s sign stands were designed to hold large display boards, he sold them, for the most part, without display boards. Sarkisian brought this suit against appellants, claiming that their distribution of the Werner device infringed his patent.
Appellants challenge the district court’s findings that the ’696 patent is valid and [674]*674that it was infringed. We will discuss each of these contentions in turn.
APPLICABLE LAW
In order for appellants to demonstrate that the trial court erred in upholding the validity of the Sarkisian patent, they must overcome the patent’s statutory presumption of validity by clear and convincing evidence. Speed Shore Corp. v. Denda, 605 F.2d 469, 471 (9th Cir. 1979). Here, that presumption has been further strengthened on appeal, since the trial court made an independent examination of the pertinent prior art in making its determination of validity.2 Those factual findings must be upheld on appeal unless they are clearly erroneous. Saf-Gard Products, Inc. v. Service Parts, Inc., 532 F.2d 1266, 1272 (9th Cir. 1976).
VALIDITY OF THE ’696 PATENT
The appellants claim that Sarkisian’s invention was not patentable under 35 U.S.C. § 103 because it was “obvious” in light of the prior pertinent art, that is, it did not produce a “synergistic” result. In our view, the district court correctly found that Sarkisian’s device was not obvious. Moreover, as we discuss in detail below, to the extent that “synergism” differs from the standard of obviousness set forth in § 103 and interpreted in Graham v. John Beere Co., 383 U.S. 1, 86 S.Ct. 684, 15 L.Ed.2d 545 (1966), we expressly disapprove it as an appropriate standard of patentability.
THE DISTRICT COURT’S ANALYSIS OF OBVIOUSNESS
35 U.S.C. § 103 establishes that a patent cannot issue if the invention would have been obvious to a person of ordinary skill in the relevant art at the time the invention was made.3 In Graham v. John Beere Co., the Supreme Court laid down a 3-part factual analysis to be followed by the federal courts in their determination of the obviousness of an invention under § 103:
Under § 103, the scope and content of the prior art are to be determined; differences between the prior art and the claims at issue are to be ascertained; and the level of ordinary skill in the pertinent art resolved. Against this background, the obviousness or non-obviousness of the subject matter is determined.
383 U.S. at 17, 86 S.Ct. at 693.
Appellants acknowledge that the obviousness or non-obviousness of an invention under § 103, though ultimately a question of law, cannot be determined without strict adherence to those factual inquiries; moreover, both the cases of this circuit4 and subsequent cases of the Supreme Court5 have reaffirmed the necessity of strict adherence to these three factual inquiries in determining the obviousness of an invention. The factual findings made by the district court pursuant to Graham, like other factual findings, must be upheld by the. appellate court unless they are clearly erroneous. Saf-Gard Products, Inc. v. Service Parts, Inc., 532 F.2d at 1272.
[675]*675The appellants argue that the district court made a number of clearly erroneous factual findings in resolving the factual issues mandated by Graham. We disagree.
1. The appellants first take issue with the district court’s assessment of the state of the “prior art” at the time Sarkisian applied for the ’696 patent. According to the appellants, the district court erroneously found that the relevant prior art was generally directed at permanently mounted signs, ignoring the fact that several of the most pertinent prior art patents related to unanchored sign stands. It is clear, however, that the court considered the fact that the prior art patents included unanchored sign stands; indeed, the four most relevant prior art patents identified by the court were for unanchored sign stands.6 The court specifically noted that a number of prior patents describe sign stands with a pedestal-type, unanchored base, and specifically outlined the features of other patents which described unanchored sign stands.7 Thus, the court’s reference to “permanently mounted signs” was decidedly not a reference to “permanently anchored sign stands.” Appellant’s contention is therefore without merit.
2. The appellants also argue that the court erred in finding that the concept of resistance to wind by “downward deflection of the display area of the sign by coil springs” was not a method previously used by manufacturers of display signs. Appellants contend that the court erroneously failed to acknowledge the many prior art patents which utilize this principle.
The court’s statement on the matter8 is somewhat ambiguous: it is not completely clear whether the court meant to say that the prior art did not contain patents utilizing the principle of downward deflection or that devices utilizing that principle were simply not manufactured by makers of display signs. The portions of the record cited by the court in support of its statement, however, indicate that the latter meaning was intended. The truth of the latter claim, as noted by the district court, is certainly supported by the record. For example, Bernard Reams, president of the largest company manufacturing metal advertising signs, testified that he had never seen a spring mounted sign stand before Sarkisian’s sign stand made its appearance.9
Appellants also argue that the court committed reversible error in failing to make specific findings on the differences between the prior art and the Sarkisian patent. Specific factual findings, while the preferable practice, are not required where the court’s entire opinion, viewed in the context of the record, shows that the district court in fact made the inquiries mandated by Graham. Palmer v. Orthokinetics, Inc., 611 F.2d 316, 319 (9th Cir. 1980). As outlined below, we find that the district court made the appropriate inquiries in reaching its conclusion that the invention was not obvious.
First, the court noted that prior art was generally directed at either permanently anchored sign stands, or at sign stands whose weighted base was essential for stability. By contrast, as the district court noted, Sarkisian’s sign has a lightweight base and is genuinely and readily portable. The district court further noted that the [676]*676approach taken by the prior art to the problem of sign stability was completely different from the approach utilized by Sarkisian. The device described in Webb’s patent used a heavy, rocking base. The device described in Mueller’s patent used a counterweight, as did the device described by Bigelow. Other signs, such as Hood, required a base weighing on the order of 75 pounds. By contrast, the unique design and interaction of parts in Sarkisian’s sign; described by the court in its opinion, enabled Sarkisian’s sign stand to remain stable under high wind conditions, regardless of the ponderousness or lightness of its base, and without the use of a counterweight. The district court’s discussion of the differences between the relevant prior art and the Sarkisian patent thus complies fully with the requirements of Graham.
SYNERGISM
Finally, appellants argue that Sarkisian’s patent should be invalidated because it is for a “combination” invention which did not produce a “synergistic” effect. According to appellants, any combination device — that is, any device composed of elements already found in the relevant prior art — must produce a “synergistic” effect in order to rise to the level of patentable invention. This “synergistic” effect is described in a number of ways by appellants: it is described, for example, as the production of “an effect greater than the sum of the several effects taken together”, as the production of “unusual and surprising” results, and as the performance of a new function by the elements in the combination. According to appellants, Sarkisian’s invention failed to produce this “synergistic” result because each element in his invention “perform[ed] the same functions as were performed outside the combination”, and alternatively, because the invention did not produce “unusual and - surprising re-suits.” In our view, however, Sarkisian’s patent must be upheld under the standard of obviousness, and therefore of patentability, for combination devices articulated in both the relevant Supreme Court decisions and in the decisions of this circuit.
Synergism and §103
The Patent Act of 1952 established that an invention must be shown to be new, useful, and non-obvious in order to be patentable. The requirement of non-obviousness, with which we are principally concerned, is set out in § 103.
Section 103, however, did not immediately bring about the uniformity in the law of patentability which its drafters had intended; the circuit courts remained split over precisely what § 103 required.10 Finally, in 1966, the Supreme Court offered the controlling interpretation of the requirements of § 103 in Graham v. John Deere Co. Recognizing that the 1952 Amendments to the Patent Act were intended to bring uniformity and definiteness to the determination of patentability, the Graham case laid down the three-part factual analysis cited earlier to be followed in the determination of non-obviousness under § 103.
The approach outlined by Graham to the analysis of non-obviousness was reaffirmed by the Court in Anderson’s-Black Rock, Inc. v. Pavement Salvage Co., 396 U.S. 57, 90 S.Ct. 305, 24 L.Ed.2d 258 (1969). In Black Rock, the Court invalidated a combination patent for an asphalt-laying device for failure to meet the standard of non-obviousness set forth in § 103, concluding that the combination in question was “reasonably obvious to one of ordinary skill in the relevant art.” In invalidating the patent, the Court admonished that “strict observance” of the requirements set forth in the Graham case [677]*677were necessary to resolve the question of obviousness. Id. at 62, 90 S.Ct. at 308. See Reeves Instrument Corp. v. Beckman Instruments, Inc., 444 F.2d 263, 271 (9th Cir. 1971). The Court’s only reference to the principle of “synergism” was a statement that “a combination of elements may result in an effect greater than the sum of the several effects taken separately. No such synergistic result is argued here.” This statement is hardly a reaffirmation of synergism as a necessary condition of patentability; it is at best either a comment on the fact that combination devices sometimes produce results which could not be produced by the components taken separately, or on the fact that combination devices occasionally- produce effects which seem greater than the sum of the effects taken separately. See Rich, Laying the Ghost of the “Invention” Requirement, in Nonobviousness— The Ultimate Condition of Patentability 1:501, 517 (1978). Republic Industries v. Schlage Lock Co., 592 F.2d 963, 970-71 (7th Cir. 1979). Neither of these is put forward by the opinion as a requirement for the patentability of combination devices.
The principles of Graham were again adhered to in Sakraida v. Ag Pro, 425 U.S. 273, 96 S.Ct. 1532, 47 L.Ed.2d 784 (1976). In Sakraida, the Court pursued the three-part inquiry mandated by Graham in concluding that the combination in question would have been obvious to a person skilled in the relevant art. The only reference made to synergism in the case was in response to the lower court’s conclusion that “synergism” was present in the device under consideration; nowhere in Sakraida does the 'court suggest that synergism is required for patentability.
Despite the holding of the Graham case and its subsequent reaffirmation in Black Rock and Sakraida, the question of the proper standard of patentability for combination inventions remains a source of constant and considerable confusion in the federal courts. Our circuit has not been completely immune from this problem; the decisions in this circuit seems to present a variegated and not always a consistent pattern of rules.11
It is admittedly difficult to reconcile completely all of the prior cases in our circuit which have addressed the question of the standard of patentability for combination patents.12 We believe, however, that [678]*678the differences among our cases are more apparent than real. In our view, the central theme of the cases remains relatively constant: Where a combination invention is involved, a searching and thorough inquiry into its obviousness may indeed be called for, appropriately reflecting the fact that patentable invention is less likely to be found in a combination of known elements. Photo Electronics Corp. v. England, 581 F.2d 772, 775 (9th Cir. 1978); see Great Atlantic & Pacific Tea Co. v. Supermarket Equipment Corp., 340 U.S. 147, 152, 71 S.Ct. 127, 130, 95 L.Ed. 162 (1950). The cases, in various ways, consistently stress this fundamental precept concerning the more intense degree of scrutiny to which combination inventions must be subjected. See, e. g., Speed Shore Corp. v. Denda, 605 F.2d 469, 471 (9th Cir. 1979); Penn International Industries v. Pennington Corp., 583 F.2d 1078, 1081 (9th Cir. 1978); Photo Electronics Corp. v. England, 581 F.2d 772, 775 (9th Cir. 1978). See M-C Industries v. Precision Dynamics Corp., 634 F.2d 1211, 1214 (9th Cir. 1980); Palmer v. Orthokinetics, Inc., 611 F.2d 316, 323 n.15 (9th Cir. 1980) [describing this more intense degree of scrutiny as a “more severe test”]. Nevertheless, the standard set out in § 103, as interpreted in Graham v. John Deere Co., and reaffirmed in Black Rock and Sakraida, remains the touchstone for determining the obviousness and hence the patentability of any invention, including combination devices.
It may be argued, of course, that the various formulations of the standard of patentability found in our cases are more than mere differences of style, but rather reflect a fundamental disagreement about the proper standard of patentability for combination devices. Synergism, in the most typical formulation, requires that an invention in some sense produce “unusual and surprising results” in order to be patentable. See Penn International Industries v. Pennington Corp., 583 F.2d 1078, 1081 (9th Cir. 1978). Those who contend that a combination device must produce a synergistic effect might argue that our cases’ continual insistence that combination devices produce “unusual or surprising results” demonstrates that “synergism” remains a requirement of patentability in this circuit.
This view that “synergism” represents an accepted standard of patentability in this circuit is based upon a fundamental misunderstanding of the sense in which a combination invention must produce “unusual or surprising results” in order to be patentable. A device may produce “unusual and surprising results” without, strictly speaking, producing “synergistic” results.13 Statements in our cases that emphasize the need for “unusual or surprising results” cannot, therefore, automatically be taken as granting an imprimatur to the synergism requirement. The sense in which “synergism” requires that a combination invention [679]*679produce “unusual or surprising results” is not the way in which our decisions require an “unusual or surprising result” for patentability, as a critical examination of our cases will demonstrate. This misunderstanding may be responsible for much of the confusion in this area.
There is a sense in which combination devices must produce “unusual or surprising” results in order to be patentable. A combination invention, by definition, draws exclusively from the pool of elements already available in the relevant prior art. The inventor’s utilization of certain elements within that pool to achieve a certain result may not, however, represent an obvious choice of elements. It may be that prior art teaches away from the use of those elements to achieve that result, that prior art is simple and the manner of combining the elements in the new device is complex, or that the elements in the new device play such different and non-obvious roles in the new device that the choice of those elements to achieve these results would not have been apparent. In each of these situations, the principle is the same: the results achievable by combining the elements selected may not have been obvious prior to 14 the time that the inventor fashioned that very combination into his or her invention.15 Where the results achievable by combining the particular elements selected were not obvious to those of ordinary skill in the relevant art prior to the time the invention was made, the combination will indeed produce “surprising” results. See Speed Shore Corp. v. Denda, 605 F.2d 469, 471 n.2 (9th Cir. 1979); Saf-Gard Products, Inc. v. Service Parts Inc., 532 F.2d at 1272.
It is important to note here that the emphasis is most properly on whether the inventor’s selection of particular elements to produce the result achieved was obvious prior to the time the invention was made. Once the combination is announced, the results may no longer seem unusual or surprising to those of ordinary skill in the relevant art. Such post-hoc perceptions, however, are not the relevant measuring stick of invention. Our cases consistently warn against the danger of concluding than an invention is “obvious” because its operative principles seem so simple and clear with the benefit of hindsight.16 See, e. g., Saf-Gard Products, Inc. v. Service Parts, Inc., 532 F.2d at 1272. As the Supreme Court noted in the case of United States v. Adams, a companion case to Graham :
It begs the question ... to state merely that magnesium and cuprous chloride were individually known battery components. If such a combination is novel, the issue is whether bringing them together as taught by the invention was obvious in light of the prior art.
383 U.S. 39, 50, 86 S.Ct. 708, 713, 15 L.Ed.2d 572 (1966). See Republic Industries Inc. v. Schlage Lock Co., 592 F.2d 963, 971 (7th Cir. 1979).
[680]*680Thus, where it is not obvious to those of ordinary skill in the relevant art prior to the time the invention was made that combining certain particular, known elements in a certain manner would achieve a certain result, the combination" which is the invention produces results which are “unusual and surprising”. The results are “unusual or surprising”, or “non-obvious” under § 103, regardless of whether the “surprising” character of the results fades once the device is revealed. This is all our cases require.17
There is a second way in which the “unusual or surprising” results requirement has been interpreted, however; though similar to that outlined above, it is far more difficult for a device to achieve. It is this second sense which seems to accord most closely with the requirement of “synergism.” Under this interpretation, the consequences of combining the elements must seem “unusual and surprising” to those of ordinary skill in the relevant art even after the elements of the combination and the manner of their combination is known. Under this interpretation, an invention is non-obvious only if claims about its successful operation are greeted with skepticism by those of ordinary skill in the relevant arts.
The fact that an invention is met with skepticism may be of assistance to a court in determining whether combining the elements to achieve the result accomplished was “obvious” to those of ordinary skill in the relevant art. Indeed, such skepticism suggests that achieving a certain result by combining particular elements in the manner done was so far removed from the typical practitioner’s perception of the state of his or her art that the achievement seemed not only not obvious, but highly improbable. This “requirement of skepticism”, however, is not a sine qua non for the patentability of combination devices. We have held a device to be equally patentable where its creation is greeted with immediate and widespread acceptance rather than with skepticism;18 indeed, widespread and immediate acceptance is itself recognized as important evidence that the device was not obvious at the time it was invented.19
Moreover, if the requirement of skepticism were a condition of non-obviousness, the caution constantly repeated in our cases against measuring the obviousness of an invention by its seeming “obviousness” after the invention is made would be meaningless, because it would always be appropriate to inquire into whether the invention seemed obvious once it was known. Thus, the requirement of synergism or of skepticism conflicts with the established case law of our circuit.20
[681]*681Synergism, moreover, also conflicts with § 103 and with the Supreme Court’s repeated caution that strict adherence to the factual inquiries in Graham be adhered to. Section 103, and the analysis required by Graham, focus on the obviousness of the invention prior to or at the time the invention was made; synergism focuses on the obviousness of the invention after it has been put together. Synergism thus turns what is nothing more than a difference in the degree of obviousness or non-obviousness into a separate and distinct measure of patentability. Such an approach is not required by logic, nor is it supported by statute or decisional law.
Other formulations of the “synergism” requirement fare no better when subjected to careful scrutiny.
The requirement that a combination device must produce a result which is “greater than the sum of its parts” is clearly only a figure of speech: an invention cannot be truly greater than the sum of its parts without creating something out of nothing, which would violate the laws of nature as we know them. The appearance of “synergism” here is simply a measure of “our imperfect knowledge of the properties of the parts,” rather than of the creation of something out of nothing. See Republic Industries v. Schlage Lock Co., 592 F.2d 963, 970 (7th Cir. 1979). Our ignorance of the properties of elements when in combination is relevant to the question of the obviousness of the combination, as discussed supra; it is not an independent test of non-obviousness.
Lastly, it is sometimes said that the elements of the combination produce a “synergistic” result if they perform a new or different function in the device in question. Again, sameness or difference of function may be of assistance to a court in determining the differences between the claimed invention and the prior art, as well as in determining the state of the prior art, both of which are relevant to the Graham analysis. However, sameness or difference of function cannot be a sufficient condition of either the obviousness or the non-obviousness of combination devices, because the elements in combination devices almost always perform functions which in some sense are the same as and in some sense are different from the roles they played in the prior art.
This seemingly inconsistent result is possible because the elements in any device do not have a single function; they play many different roles at the same time. On the most general level of description, for example, the elements in a combination device will play the “same” role in the invention as they play in every other mechanical device of which they are a part: as Judge Learned Hand once noted, “all machines are made up of the same elements: rods, pawls, pit-mans, journals, toggles, gears, cams and the like, all acting their parts as they always do and always must.”21 Even when we narrow our focus to the specific role played by the elements in question in the prior pertinent art, it is almost inevitable that the elements under scrutiny will play “the same” role in the new invention as they played in the devices of prior art: by definition the elements composing the new “combination” invention are already found in and are taken from the prior art. Finally, it is almost inevitable that the elements of the invention will play “a new and different role” in the new invention than they played in the prior art, because the “function” of any element can reasonably be described in terms of the role that element plays in achieving the new result brought about by the invention.
[682]*682The case before us provides an illustration of the danger of focusing too narrowly on the question of whether the elements of the invention under scrutiny perform a “new and different” function. One of the crucial elements in Sarkisian’s device is its extension springs. With all tension springs, including those utilized in the prior art, the extension springs in Sarkisian’s device play the “same” function of remaining tightly wound unless separated by an external force. More specifically, the extension springs in Sarkisian’s device share with those of many prior art references the function of holding the sign stand’s display board in an upright position. However, if the springs’ function is described in terms of their function in the device as a whole, or in terms of the role they play in achieving the novel result obtained by Sarkisian’s device, the springs suddenly take on a “new and different function:” they enable the display board to deflect without toppling the entire sign stand over. Thus, depending on how one chooses to describe the role of the extension springs, they play both “the same function” and a “new and different function” from those in prior art devices.
Any combination invention which introduces some feature of novelty — in its results, in its manner of operation, or in any other significant way — will utilize elements which can be described as performing a “new or different function.” The element’s “function” need only be described in terms of its role in contributing to the invention’s novelty to achieve that result. At the same time, it will almost always be the case that these elements play the “same” role in the invention as they played in the prior art from which they are, by definition, taken. It cannot be, therefore, that a combination invention is non-obvious whenever its elements perform a “new and different function”; nor is it true that “sameness” of function is by itself a definitive indication of obviousness. Sameness and difference in function are simply factors to be considered in answering the question of obviousness; the basic question remains whether any difference in function is significant enough to support the conclusion that the device in question was not obvious to those of ordinary skill in the relevant art.
Sarkisian’s In ven tion
There is ample support in the record for the district court’s conclusion that Sarkisian’s device was not obvious to one of ordinary skill in the relevant art. The court could reasonably conclude that nothing in the prior art suggested that combining a pair of extension springs, a frame, a lightweight base and a sign of certain dimensions could result in an unanchored, truly portable, yet stable sign stand. Indeed, appellant Werner described his own stand, which utilized the same principles as Sarkisian’s patent, as “amazing” and as a “breakthrough in the construction sign field.”
Prior art, of course, did contain the elements of Sarkisian’s device. Contrary to appellants’ assertion, however, these elements did not perform functions in the pri- or art which were similar or even analogous to the functions played by the elements in Sarkisian’s device. For example, “spaced apart” legs, similar to those described in Sarkisian’s ’696 patent, are illustrated in the base of Thompson. However, in Thompson the function of the “spaced apart legs” was simply to provide a secure base for placement of a sign. There is no evidence that the “spaced apart” structure of the base in Thompson had any role in the device’s ability to withstand an impinging force — indeed, as the district court noted, there is no evidence that Thompson’s sign stand was designed to be wind proof at all. Similarly, the use of “spring means” both to impart flexibility to a sign stand and to return the display sign to a vertical position following removal of a deflecting force is shown by a substantial number of previous patents. In some of these patents,22 the base of the device is either permanently [683]*683affixed to the ground or is inserted into a holder that is permanently attached. In others, the spring means is attached to a heavy, unanchored base.23 The primary determinant of the stability of these sign stands is the sign’s base, which is either permanently affixed to the ground or is made of extremely heavy material in relation to the rest of the device.24 The genuinely novel function performed by the extension springs in Sarkisian’s device was an important fact supporting the district court’s finding of non-obviousness.
Moreover, in none of the prior patents must the amount of tension in the springs be correlated to the other features of the sign stand in order for the sign stand to be stable, or for the springs to perform their function in the device properly. In Sarkisian’s device, by contrast, the amount of tension in springs is crucial to the sign stand’s stability.
There are also patents in the prior art that illustrate the use of pairs of non-coneentric or “spaced apart” springs within the meaning of “spaced apart” in Sarkisian’s patent.25 While the spring pairs in all of these devices, including Sarkisian’s, prevent the sign from “canting” or twisting about its vertical axis, the primary function of the “spaced-apart” springs in Sarkisian’s device is to control the direction in which a sign stand is deflected when struck by a vehicle, away from the line of traffic. No similar function is even required in the prior art patents cited, since all of these prior art patents describe devices which are permanently anchored.
In short, then, while elements of Sarkisian’s device appear in the prior art, none of the prior art devices had suggested combining these elements in the unique fashion done by Sarkisian to achieve the results he achieved. The prior art suggested that road signs required either a heavy or permanently anchored base in order to be stable, that the spring means played a relatively minor role in contributing to the stability of unanchored sign stands, and that the precise amount of initial tension in the extension springs was not a significant factor in enabling portable sign stands to withstand an impinging force without tipping over. Sarkisian’s combination of elements found in the prior art thus produced a result that was unusual and surprising, because it produced a readily portable, spring mounted, windproof sign stand with a lightweight base.
The trial court’s conclusion that Sarkisian’s invention was not obvious under § 103 is thus fully supported by the evidence of record.26
The trial court also properly considered so-called “secondary” indicia of non-obviousness in upholding the validity of Sarkisian’s patent. These considerations, such as the commercial success of the device, long-[684]*684felt but unsolved needs, and prior unsuccessful attempts to solve the problem ultimately addressed by the new invention are valid and relevant criteria in determining the level of ordinary skill in the relevant art, and therefore in determining obviousness.27 Palmer v. Orthokinetics, Inc., 611 F.2d 316, 324-5 (9th Cir. 1980). Santa Fe-Pomeroy, Inc., v. P & Z, Co., 569 F.2d 1084, 1097-98 (9th Cir. 1978). The court’s finding that Sarkisian’s sign has met with substantial commercial success is fully supported by the evidence; appellants apparently do not contest this fact. Similarly, the court’s finding that the Sarkisian sign solved the longstanding problem of such sign stands walking and tipping in the wind is fully supported by the evidence. The evidence also indicated that no satisfactory “readily portable” road signs were available before Sarkisian’s invention; testimony indicated that the existing unanchored road signs had to be weighted down with 200 to 300 pounds of sandbags in order to maintain their stability. These secondary considerations also support the trial court’s conclusion that Sarkisian’s device was not obvious. We find no error.
INFRINGEMENT
A. Infringement
The district court found that appellants’ device induced infringement of the claims of the ’696 patent.28 While finding that appellants’ devices did not come within the literal language of the patent’s claims, the court held that the differences which existed between appellant’s sign and the Sarkisian patent were insignificant, and that the accused devices were the equivalents of the device claimed in the ’696 patent. We affirm.29
The language of a patent’s claims provide the starting point for determining the scope of the patent’s monopoly, and initially provides the measure against which infringement is determined. In a combination patent, the patent is infringed only if the accused device contains all of the elements set forth in the patent’s claims. Nelson v. Batson, 322 F.2d 132, 135 (9th Cir. 1963).
The literal language of the patent’s claims, however, is not the sole measure of infringement. A patent might be rendered a “hollow and useless thing” were its scope confined to the literal language of its claims, since the making of a minor and otherwise insignificant variation in the invention might then take a pirating device outside the claim’s literal language and hence outside the protection of the patent. The doctrine of equivalents has evolved in response to this problem. Under that doctrine, two devices are considered to be the same if they do substantially the same work in substantially the same manner, and accomplish substantially the same results. Union Paper Bag Machine Co. v. Murphy, 97 U.S. 120, 125, 24 L.Ed. 935 (1878). The determination of infringement is therefore [685]*685a complex inquiry: even where elements of an accused device fall outside the literal language of a patent’s claims, the court must determine whether those elements are the “equivalents” of elements set forth in the patent’s claims in order to determine whether infringement has occurred.
Whether an element in an accused device is the “equivalent” of an element described in a patent may turn on how expansively or narrowly the relevant language in the patent is construed. The range of equivalents accorded a patent is in large part a function of the importance and originality of the invention in question. Since the purpose of the doctrine of equivalents is to secure for the inventor a just reward for his or her invention, patents which represent important and significant advances in the relevant technology are entitled to a broad range of equivalents; similarly, patents which represent a rather small advance in a crowded field are entitled to a correspondingly narrow range of equivalents. Nelson v. Batson, 322 F.2d 132, 135 (9th Gir. 1963).
The district court’s finding of infringement rested primarily on its factual determination that two essential features of the accused devices were the equivalents of elements described in the claims of the ’696 patent. First, the court found that the design of the legs constituting the base of the accused signs were the equivalents of the “generally parallel” legs described in the claims of the ’696 patent. The court also found that even though the pair of springs in the accused devices were in physical contact with each other they were the equivalent of the “spaced apart” springs described in the claims of the ’696 patent.30
The appellants contend that the range of equivalents to which Sarkisian’s patent is entitled is so narrow that it cannot encompass the features of the accused devices just described. According to appellants, Sarkisian’s sign stand represented at best a minor advance in a crowded and highly developed field. While the evidence indicates that numerous prior patents existed for both anchored and unanchored sign stands, the evidence also fully supports the district court’s assumption that Sarkisian’s device represented a significant advance in the field of unanchored, readily portable, windproof sign stands.31 As discussed in Part I of this opinion, the record fully supports the trial court’s factual determination that Sarkisian’s device was a significant step in the progress of the relevant art and was therefore entitled to a range of equivalents broad enough to encompass the appellants’ device.
1. The record fully supports the district court’s determination that the base structure of the appellants’ latest sign stand model was the equivalent of the base structure of the sign stand described in the ’696 patent. As illustrated by Figure 6, the legs of the accused device are not “generally parallel” to each other as described in the claims of Sarkisian’s ’696 patent. However, as appellee’s expert testified, the legs in the accused device and in Sarkisian’s device perform precisely the same function, which is to stabilize both the base and the sign stand as a whole. Moreover, the record supports the conclusion that this function is performed in precisely the same fashion in both devices. So long as the base is of a certain length and a certain minimum weight, precisely correlated with other dimensions and features of the sign stand and its springs, and so long as the stand’s center of gravity is within the base, the base will stabilize the sign stand and remain firmly planted even under extremely high wind conditions. Both the accused devices and Sarkisian’s device utilize this principle in fashioning the bases of their respective de[686]*686vices. The fact that the legs of the accused device are not parallel to each other does not change the fact that the base of that device embodies the same principle used in the Sarkisian patent and functions in precisely the same way in achieving the same result. The district court’s finding of equivalence is therefore not clearly erroneous.
2. The district court also found that the pair of springs used in the accused sign stand was the equivalent of the pair of “spaced apart” coil springs described in the claims of Sarkisian’s ’696 patent. The pair of springs in the Sarkisian patent served two functions: they minimized canting or twisting of the display board around its vertical axis, and they assured that the sign stand would be deflected away from the line of traffic should the stand be struck by a vehicle. The evidence established that the crucial determinant of the spring’s effectiveness for these purposes is not the horizontal distance between the springs, but is whether the springs are located on the same vertical axis. So long as the springs are located on different vertical axes, they will function effectively regardless of how close or how “spaced apart” they are on the horizontal plane. The pair of springs on the accused device, located on different vertical axes, function in precisely the same way as the springs on Sarkisian’s device, achieving the same results in precisely the same manner. The district court’s finding of equivalence is therefore fully supported by the evidence and is not clearly erroneous.
B. Contributory Infringement
Finally, the record supports the court’s determination that appellants’ sign stand induced infringement of the ’696 patent. While most of the accused sign stands were sold without a display board with a “relatively large surface area,” they are specifically designed to hold such signs. Appellant Werner acknowledged as much in his deposition. We uphold the district court’s finding of contributory infringement. 35 U.S.C. § 271(c).32
DOUBLE PATENTING
Sarkisian cross-appeals from the district court’s judgment invalidating portions of claims 1 and 2 of the ’482 patent on grounds of double patenting. For the reasons stated below, we reverse the district court’s finding of invalidity.
Sarkisian filed the ’482 patent application on July 30, 1970, while the ’696 patent application was still pending.33 The ’482 patent issued some two months after the ’696 patent, accompanied by a “terminal disclaimer” which dedicated to the public the portion of ’482’s term which extended beyond the term of the ’696 patent.34
[687]*687The claims of the ’482 patent in question here, claims 1 and 2,35 overlap considerably with claims of the ’696 patent. The claims of the ’482 patent are, however, somewhat broader; the spring mechanism in the ’482 patent is claimed as a pair of “spring means,” rather than as a pair of “coil springs” in compression, as in the ’696 patent. Similarly, the legs of the ’482 patent are described simply as “ground engaging means,” eliminating the limitation of “generally parallel” legs found in the ’696 patent36
The district court held that claims 1 and 2 of the ’482 patent claimed “the same” device as that described in the claims of the ’696 patent to the extent that the claims of the two patents overlapped. According to the district court, the ’482 patent thus represented an impermissible attempt to obtain two patents on the same invention, in violation of 35 U.S.C. § 102. The court invalidated those portions of claims 1 and 2 which overlapped with claims in the ’696 patent. In our view, the court’s analysis in this difficult area of patent law was incorrect.
The potential double patenting problem typically arises when an inventor seeks multiple patents which relate in some way to the same invention. Double patenting problems fall into one of two classifications: “same invention” double patenting, and “obvious” type double patenting, sometimes referred to as “extension of monopoly” double patenting.
[688]*688“Same invention” double patenting arises where an inventor seeks to obtain two patents on the same invention. “Same invention” double patenting runs afoul of the stricture of 35 U.S.C. § 102, that a single patent shall issue for a single invention. Where, on the other hand, the invention claimed in a later patent application is not “the same” as that claimed by the first or “parent” application, but is rather an “obvious variation” of that claimed in the parent application, “obviousness” type double patenting is involved.37
The major objection to “obviousness” type double patenting, though certainly not the only one,38 is that the granting of multiple patents may result in an unwarranted extension of the original patent’s monopoly beyond the time period allowed by law. The current statutory scheme rewards the inventor for disclosing and making the invention with a 17-year monopoly on his or her invention, during which the patentee can exclude others from making, using, or selling the device.39 Where a subsequent application disclosing an obvious variation of the initial invention is granted after the parent application has been granted, the original patent’s 17-year monopoly is impermissibly extended until the subsequent patent expires, denying the public its statutory right to use the invention at the end of the first monopoly period. See In re Robeson, 331 F.2d 610, 614, 51 C.C.P.A. 1271 (1964).
The courts wrestled for some time with the myriad problems, both metaphysical and semantic, involved in distinguishing patent claims which were “the same” from those which were merely “similar to” or “colorable variations of” the claims in the first patent application.40 Finally, the Court of Custom and Patent Appeals, in In re Vogel, 422 F.2d 438, 57 C.C.P.A. 920 (1970) set out a two-step process, guided by specific, workable criteria, for analyzing a potential double patenting problem:
A good test, and probably the only objective test, for “same invention,” is whether one of the claims could be literally infringed without literally infringing the other. If it could be, the claims do not define identically the same invention. .. .
If it is determined that the same invention is being claimed twice, 35 U.S.C. § 101 forbids the grant of the second patent, regardless of the presence or absence of a terminal disclaimer. If the same invention is not being claimed twice, a second question must be asked.
The second analysis question is: Does any claim in the application define merely an obvious variation of an invention disclosed and claimed in the patent? ...
If the answer to the second question is no, there is no double patenting involved .... If the answer is yes, a terminal disclaimer is required to prevent undue time timewise extension of monopoly.
In re Vogel, 422 F.2d at 441-42. Thus, under Vogel, the claims of the parent application must first be compared with the claims of the later application to determine whether the claims of either application could be literally infringed without literally infringing the claims of the other. Pursche v. Atlas Scraper & Engineering Co., 300 [689]*689F.2d 467, 479-80 (9th Cir.), cert. denied, 371 U.S. 911, 83 S.Ct. 251, 9 L.Ed.2d 170 (1962).41 If so, the claims at issue do not define the “same invention,” and, as to those claims, “same invention” double patenting is not involved. If “same invention” double patenting is not involved, it must then be determined whether the claims of the subsequent application define an “obvious variation” of the invention claimed in the parent application. If they do, then “obviousness” type double patenting is made out.
The district court correctly concluded that the claims in the ’482 patent are broader than the claims of the ’696 patent, and that devices which might infringe the literal terms of the ’482 patent might not infringe those of the narrower ’696 patent. The court erroneously concluded, however, that because a portion of each claim in the ’482 patent overlapped with some claim in the ’696 patent, same invention double patenting was involved “to some extent.” We believe this conclusion to be incorrect. In order to determine whether the invention defined by claims in the subsequent application was the “same” as the invention claimed in the parent application, the entire claim in each application must be looked to, since it is the claim taken as a totality which defines the invention.42
Under the test set forth in Vogel, obviousness type double patenting is clearly involved in this case. Claims 1 and 2 of the ’482 application, viewed in their entirety, are clearly broader than the claims in the ’696 patent, and hence do not define “the same” invention. The question then becomes whether the ’482 patent should be invalidated for “obviousness” type double patenting.
It appears to us, and the parties do not suggest otherwise, that the sole objection to double patenting raised by the present case is the danger that the ’482 patent may impermissibly extend the monopoly of the ’696 patent. The issue is thus whether the terminal disclaimer filed by Sarkisian as to the ’482 patent obviates that problem. We conclude that it does.
This circuit has previously indicated that subsequent patent applications or patents are not void for “obviousness” type double patenting so long as there is no danger that the later application or patent will impermissibly extend the original patent’s monopoly. Pursche v. Atlas Scraper & Engineering Co., 300 F.2d at 480-81. While Pursche involved a series of patents which simply expired on the same day, without a terminal disclaimer, the principle articulated by that case is fully applicable here: “. . . the rule [voiding later patents for double patenting] will not operate where the reason for its application is absent.” Pursche, 300 F.2d at 481.
We find that the terminal disclaimer effectively dispensed with the problem of the ’482 patent’s extending the monopoly of the ’696 patent beyond the time permitted by statute, and we reverse the lower court’s finding of invalidity.
AFFIRMED in part, REVERSED in part.