Mitchell, J.

The plaintiff being the owner of a quantity of logs marked “H. C. A.,” cut in the winters of 1882 and 1883, and lying in the Mississippi river, or on its banks, ábove Minneapolis, defendant and the plaintiff, through his agent, D. S. Mooers, having fully agreed on the térras of a sale and purchase of the logs referred to, executed the following written agreement;

‘‘■‘agreement.

''“-Hastings, Minn., June 1, 1883.

“I have this day sold to E. C. Libby, of Hastings, Minn., all my logs marked ‘ H. C. A.,’ cut in the winters of 1882 and 1883, for ten dollars a thousand feet, boom scale at Minneapolis, Minnesota. Payments cash as fast as scale bills are produced.

[Signed] “J. H. Thompson,

“Per D. S. Mooers.

"B. C. Libby.”

This action having been brought for the purchase-money, the defendant — having pleaded a warranty of the quality of the logs, alleged to have been made at the time of the sale, and a breach of it —offered on the trial oral testimony to prove the warranty, which •was admitted, over the objection of plaintiff that it was incompetent to prove a verbal warranty, the contract of sale being in writing. This raises the only point in the case.

No ground was laid for the reformation of the written contract, and any charge of fraud on part of plaintiff or his agent in making *377the sale was on the trial expressly disclaimed. No rule is more familiar than that “parol contemporaneous evidence is inadmissible to contradict or vary the terms of a valid written instrument,” and yet none has given rise to more misapprehension as to its application. It is a rule founded on the obvious inconvenience and injustice that would result if matters in writing, made with consideration and deliberation, and intended to embody the entire agreement of the parties, were liable to be controlled by what Lord Coke expressively calls “the uncertain testimony of slippery memory.” Hence, where the parties have deliberately put their engagements into writing in such terms as to import a legal obligation, without any uncertainty as to the object or extent of such engagement, it is conclusively presumed that the whole engagement of the parties, and the manner and extent of their undertaking, was reduced to writing.^ 1 Greenl. Ev. § 275. Of course, the rule presupposes that the parties intended to have the terms of their complete agreement embraced in the writing and hence it does not apply where the writing is incomplete on its face and does not purport to contain the whole agreement, as in the case of mere bills of parcels, and the like.

But in what manner shall it be ascertained whether the parties intended to express the whole of their agreement in writing? It is sometimes loosely stated that where the whole contract be not reduced to writing, parol evidence may be admitted to prove the part omitted. But to allow a party to lay the foundation for such parol evidence by oral testimony that only part of the agreement was reduced to writing, and then prove by parol the part omitted, would be^to work in a circle, and to permit the very evil which the rule was designed to prevent. The only criterion of the completeness of the written con- j tract as a full expression of the agreement of the parties is the writing itself. it imports on its face to be a complete expression of f the whole agreement, — that is, contains such language as imports a j complete legal obligation, — it is to be presumed that the parties have j introduced into it every material item and term; and parol evidence cannot be admitted to add another term to the agreement, although the writing contains nothing on the particular one to which the parol evidence is directed. The rule forbids to add by parol where the *378writing is silent, as well as to vary where it speaks, — 2 Phil. Evidence, (Cow. & H. Notes,) 669; Naumberg v. Young, 44 N. J. Law, 331; Hei v. Heller, 53 Wis. 415, — and the law controlling the operation of a written contract becomes a part of it, and cannot be varied by parol any more than what is written. 2 Phil. Ev. (Cow. & H. Notes,) 668; La Farge v. Rickert, 5 Wend. 187; Creery v. Holly, 14 Wend. 26; Stone v. Harmon, 31 Minn. 512.

The written agreement in the case at bar, as it appears on its face, in connection with the law controlling its constru'ction and operation, purports to be a complete expression of the whole agreement of the parties as to the sale and purchase of these logs, solemnly executed by both parties. There is nothing on its face (and this is a question of law for the court) to indicate that it is a mere informal and incomplete memorandum. Parol evidence of extrinsic facts and eircum-' stances would, if necessary, be admissible, as it always is, to apply the contract to its subject-matter, or in order to a more perfect understanding of its language. But in that case such evidence is used, not to contradict or vary the written instrument, but to aid, uphold, and enforce it as it stands. The language of this contract “imports a legal obligation, without any uncertainty as to its object or the extent of the engagement,” and therefore “it must be conclusively presumed that the whole engagement of the parties, and the manner and extent of the undertaking, was reduced to writing.” No new term, forming a mere incident to or part of the contract of sale, can be added by parol.

That in case of a sale of personal property a warranty of its quality is an item and term of the contract of sale, and not a separate and independent collateral contract, and therefore cannot be added to the written agreement by oral testimony, has been distinctly held by this court, in accordance, not only with the great weight of authority, but also, as we believe, with the soundest principles. Jones v. Alley, 17 Minn. 269, (292.)

We are referred to Healy v. Young, 21 Minn. 389, as overruling this. This is an entire misapprehension of the point decided in the latter case. In Healy v. Young the claim of defendant was that for a certain consideration plaintiff agreed verbally to release a certain *379debt, and also to convey certain personal property; and that, in part-performance of this prior verbal agreement, he executed a bill of sale of the property. What was decided was that the execution in writing of the bill of sale in part-performance of this verbal agreement did not preclude defendant from proving by parol the prior agreement. The parties had not put their original agreement in writing, and the bill of sale executed in part-performance in no way superseded it. Moreover, the promise to release the debt was a distinct collateral matter from that covered by the bill of sale, and in that view of the ease it was immaterial whether the oral agreement preceded or was contemporaneous with the bill of sale.

In opposition to the doctrine of Jones v. Alley, we are referred to a few cases which seem to hold that parol evidence of a warranty is admissible on the ground that a warranty is collateral to the contract of sale, and that the rule does not exclude parol evidence of matters collateral to the subject of the written agreement. It seems to ns that this is based upon a misapprehension as to the sense in which the term “collateral” is- used in the rule invoked. There are a great many matters that, in a general sense, may be considered collateral to the contract; for example, in the case of leases, covenants for repairs, improvements, payment of taxes, etc., are, in a sense, collateral to a demise of the premises. But parol evidence of these would not be admissible to add to the terms of a written lease. So, in a sense, a warranty is collateral to a contract of sale, for the title would pass without a warranty. It is also collateral in the sense that its breach is no ground for a rescission of the contract by the vend: but that he must resort to his action on the warranty for damages/ But, when made, a warranty is a part of the contract of sale. The common sense of men would say, and correctly so, that when, on a sale of personal property, a warranty is given, it is one of the terms of the sale, and not a separate and independent contract. To justify the admission of a parol promise by one of the parties to a written contract, on the ground that it is collateral, the promise must relate to a subject distinct from that to which the writing relates. Dutton v. Gerrish, 9 Cush. 89; Naumberg v. Young, supra; 2 Taylor, Ev. § 1038. See Lindley v. Lacey, 34 Law J., C. P., 7.

*380We have carefully examined all the cases cited in the quite exhaustive brief of counsel for defendant, and find but very few that are 'at all in conflict with the views already expressed, and these few do not commend themselves to our judgment. Our conclusion therefore is that the court erred in admitting parol evidence of a warranty, and therefore the order refusing a new trial must be reversed.

OPINION

Memorandum Decision of the Court of Appeals, Division One, filed September 17, 1991, vacated and remanded

FELDMAN, Chief Justice.

Bobby Sid Taylor petitions us to review a decision reversing a jury verdict in his fa*151vor in a bad faith claim against State Farm Mutual Automobile Insurance Co. He argues that the court of appeals erroneously held that his bad faith claim was barred by a release he signed in 1981. We granted review because the case raises important issues in the area of contract and insurance law. We have jurisdiction pursuant to Ariz. Const, art. VI, § 5(3), and Ariz.R.Civ. App.P. 23.

FACTS AND PROCEDURAL HISTORY

This insurance bad faith action arises out of an accident that occurred approximately sixteen years ago. Many of the facts are undisputed. The accident involved three vehicles — one occupied by Anne Ring and passenger James Rivers, the second by Douglas Wistrom, and the third by Bobby Sid Taylor. Ring, Rivers, and Taylor all were injured. The facts surrounding the accident are set forth in Ring v. Taylor, 141 Ariz. 56, 59, 685 P.2d 121, 124 (Ct.App.1984). Ring, her husband, and Rivers filed actions against Taylor and Wistrom. These actions were consolidated before trial. Taylor’s insurer, State Farm, retained attorney Leroy W. Hofmann to defend Taylor. Taylor also personally retained attorney Norman Bruce Randall, who filed a counterclaim against Ring for Taylor’s damages. Taylor, therefore, was represented by both Randall and Hofmann in the matter. Because the Rings and Rivers agreed with Wistrom to a stipulated judgment and covenant not to execute, Taylor was the only party vulnerable to the Ring/Rivers claims. At trial, the Rings and Rivers obtained combined verdicts against Taylor for approximately $2.5 million in excess of his insurance policy limits. The court of appeals affirmed these judgments. Taylor, 141 Ariz. at 59, 71, 685 P.2d at 124, 136.

The Rings eventually settled with State Farm. Taylor, however, sued State Farm for bad faith seeking damages for the excess Rivers judgment, claiming, among other things, that State Farm improperly failed to settle the Rivers matter within policy limits. State Farm moved for summary judgment, asserting that Taylor relinquished his bad faith claim when, in 1981, he signed a release drafted by attorney Randall in exchange for State Farm’s payment of $15,000 in uninsured motorist benefits.1 Taylor also moved for partial summary judgment, seeking a ruling that, as a matter of law, the release did not preclude his bad faith claim. The judge denied both motions, finding that the release was ambiguous and that therefore parol evidence was admissible at trial to aid in interpreting the release. A second judge, who presided at trial, also denied State Farm’s motion for directed verdict based on the release. Having been instructed on the interpretation of the release, the jury returned a verdict in favor of Taylor for compensatory damages of $2.1 million. The court also awarded Taylor $300,000 in attorney fees.

The court of appeals reversed, holding that the release agreement was not ambiguous and therefore the judge erred by admitting parol evidence to vary its terms. Taylor v. State Farm Mut. Auto. Ins. Co., No. 1 CA-CV 9908 (Sep. 17, 1991) (mem. dec.). Based on the agreement’s “four corners,” the court held that “it clearly release[d] all policy contract rights, claims, and causes of action that Taylor has or may have against State Farm.” Id. at 20. According to the court, because the release should have been strictly enforced, there was no basis for Taylor’s bad faith claim. Id. We believe the court’s decision both incorrectly applies settled legal principles and raises unsettled issues of contract interpretation.

DISCUSSION

Much of the dispute in this case centers on the events that surround the drafting of the release and the inferences that can be drawn from those events. As noted, the trial court found that the release was am*152biguous and admitted extrinsic evidence to aid in its interpretation. The court of appeals found no ambiguity. Taylor, mem. dec. at 20, 23. In resolving this issue, we must address the scope and application of the parol evidence rule in Arizona and decide whether, under these facts, the trial court properly admitted extrinsic evidence to interpret the release.

A. Legal principles

The application of the parol evidence rule has been the subject of much controversy and scholarly debate. See generally Darner Motor Sales, Inc. v. Universal Underwriters Ins. Co., 140 Ariz. 383, 392-93, 682 P.2d 388, 397-98 (1984); John D. Calamari & Joseph M. Perillo, The Law of Contracts §§ 3-2 to 3-16, at 135-77 (3d ed. 1987); Robert L. Gottsfield, Darner Motor Sales v. Universal Underwriters: Corbin, Williston and the Continued Viability of the Parol Evidence Rule in Arizona, 25 Ariz.St.L.J. 377 (1993). “When two parties have made a contract and have expressed it in a writing to which they have both assented as the complete and accurate integration of that contract, evidence, whether parol or otherwise, of antecedent understandings and negotiations will not be admitted for the purpose of varying or contradicting the writing.” 3 Arthur L. Corbin, Corbin on Contracts § 573, at 357 (1960) (“Corbin”); see also Rental Dev. Corp. v. Rubenstein Const. Co., 96 Ariz. 133, 136, 393 P.2d 144, 146 (1964) (citing Corbin). Antecedent understandings and negotiations may be admissible, however, for purposes other than varying or contradicting a final agreement. 3 Corbin § 576, at 384. Interpretation is one such purpose. 3 Corbin § 579, at 412-13; Restatement (Second) of Contracts § 214(c) & cmt. b (1979) (“Restatement”).

Interpretation is the process by which we determine the meaning of words in a contract. See Restatement § 200. Generally, and in Arizona, a court will attempt to enforce a contract according to the parties’ intent. See Darner, 140 Ariz. at 393, 682 P.2d at 398; Polk v. Koerner, 111 Ariz. 493, 495, 533 P.2d 660, 662 (1975); Sam Levitz Furniture Co. v. Safeway Stores, Inc., 105 Ariz. 329, 330-31, 464 P.2d 612, 613-14 (1970). “The primary and ultimate purpose of interpretation” is to discover that intent and to make it effective. 3 Corbin § 572B, at 421 (1992 Supp.). The court must decide what evidence, other than the writing, is admissible in the interpretation process, bearing in mind that the parol evidence rule prohibits extrinsic evidence to vary or contradict, but not to interpret, the agreement. See 3 Corbin § 543, at 130-34. These substantive principles are clear, but their application has been troublesome.

1. Restrictive view

Under the restrictive "plain meaning” view of the parol evidence rule, evidence of prior negotiations may be used for interpretation only upon a finding that some language in the contract is unclear, ambiguous, or vague. E. Allan Farnsworth, Farnsworth on Contracts § 7.12, at 270 (1990) (“Farnsworth”). Under this approach, “if a writing, or the term in question, appears to be plain and unambiguous on its face, its meaning must be determined from the four corners of the instrument without resort to extrinsic evidence of any nature.” Calamari & Perillo, supra § 3-10, at 166-67; cf. Gottsfield, supra, at 388-89. Thus, if the judge finds from the face of a document that it conveys only one meaning, parol evidence is neither considered nor admitted for any purpose. The danger here, of course, is that what appears plain and clear to one judge may not be so plain to another (as in this case), and the judge’s decision, uninformed by context, may not reflect the intent of the parties.

2. Corbin view

Under the view embraced by Professor Corbin and the Second Restatement, there is no need to make a preliminary finding of ambiguity before the judge considers extrinsic evidence. 3 Corbin § 542, at 100-05 (1992 Supp.); Restatement § 212 cmt. b; Farnsworth § 7.12, at 272; Gottsfield, supra, at 384. Instead, the *153court considers all of the proffered evidence to determine its relevance to the parties’ intent and then applies the parol evidence rule to exclude from the fact finder’s consideration only the evidence that contradicts or varies the meaning of the agreement. 3 Corbin § 542, at 100-01 (1992 Supp.). According to Corbin, the court cannot apply the parol evidence rule without first understanding the meaning the parties intended to give the agreement. Id. To understand the agreement, the judge cannot be restricted to the four corners of the document. Again, even under the Corbin view, the court can admit evidence for interpretation but must stop short of contradiction. See 3 Corbin § 574, at 371-72; Gottsfield, supra, at 386-87, 392.

3. Arizona view

Writing for a unanimous court in Smith v. Melson, Inc., 135 Ariz. 119, 121-22, 659 P.2d 1264, 1266-67 (1983), Chief Justice Holohan expressly committed Arizona to the Corbin view of contract interpretation. Burkons v. Ticor Title Ins. Co. of Cal., 168 Ariz. 345, 350-51, 813 P.2d 710, 715-16 (1991); see also Darner, 140 Ariz. at 393, 682 P.2d at 398; cf. Gottsfield, supra, at 389-90 & n. 83 (citing numerous cases decided before Melson). We have not, however, fully explored Melson’s application. See Gottsfield, supra, at 378. We have held that a court may consider surrounding circumstances, including negotiation, prior understandings, and subsequent conduct, but have not elaborated much further. Darner, 140 Ariz. at 393, 682 P.2d at 398; see also Burkons, 168 Ariz. at 351, 813 P.2d at 716; Melson, 135 Ariz. at 122, 659 P.2d at 1267.

According to Corbin, the proper analysis has two steps. First, the court considers the evidence that is alleged to determine the extent of integration, illuminate the meaning of the contract language, or demonstrate the parties’ intent. See 3 Corbin § 542, at 100-01 (1992 Supp.). The court’s function at this stage is to eliminate the evidence that has no probative value in determining the parties’ intent. Id. The second step involves “finalizing” the court’s understanding of the contract. Id. at 100. Here, the parol evidence rule applies and precludes admission of the extrinsic evidence that would vary or contradict the meaning of the written words. Id.

Even during the first step, the judge may properly decide not to consider certain offered evidence because it does not aid in interpretation but, instead, varies or contradicts the written words. See id. at 101. This might occur when the court decides that the asserted meaning of the contract language is so unreasonable or extraordinary that it is improbable that the parties actually subscribed to the interpretation asserted by the proponent of the extrinsic evidence. See id. “The more bizarre and unusual an asserted interpretation is, the more convincing must be the testimony that supports it.” 3 Corbin § 579, at 420. At what point a judge stops “listening to testimony that white is black and that a dollar is fifty cents is a matter for sound judicial discretion and common sense.” Id.

When interpreting a contract, nevertheless, it is fundamental that a court attempt to “ascertain and give effect to the intention of the parties at the time the contract was made if at all possible.” Polk, 111 Ariz. at 495, 533 P.2d at 662; see also Darner, 140 Ariz. at 393, 682 P.2d at 398; Sam Levitz Furniture Co., 105 Ariz. at 330-31, 464 P.2d at 613-14. If, for example, parties use language that is mutually intended to have a special meaning, and that meaning is proved by credible evidence, a court is obligated to enforce the agreement according to the parties’ intent, even if the language ordinarily might mean something different. See Restatement § 212 cmt. b, illus. 3 & 4. The judge, therefore, must avoid the often irresistible temptation to automatically interpret contract language as he or she would understand the words. This natural tendency is sometimes disguised in the judge’s ruling that contract language is “unambiguous.” See 3 Corbin § 543A, at 159 (1992 Supp.). Words, however, are seldom so clear that they “apply themselves to the subject matter.” Restatement § 214 cmt. b. On occasion, exposition of the evidence regarding *154the intention of the parties will illuminate plausible interpretations other than the one that is facially obvious to the judge. See id. Thus, ambiguity determined by the judge’s view of “clear meaning” is a troublesome concept that often obstructs the court’s proper and primary function in this area — to enforce the meaning intended by the contracting parties. See 3 Corbin § 542, at 122-24; Gottsfield, supra, at 385.

Recognizing these problems, we are hesitant to endorse, without explanation, the often repeated and usually oversimplified construct that ambiguity must exist before parol evidence is admissible. We have previously criticized the ambiguity prerequisite in the context of non-negotiated agreements. See State Farm Mut. Auto. Ins. Co. v. Wilson, 162 Ariz. 251, 257, 782 P.2d 727, 733 (1989) (recognizing the lack of logic in requiring ambiguity, which may be fortuitous, to prove the true terms of an agreement); Darner, 140 Ariz. at 389, 682 P.2d at 394 (same). Moreover, a contract may be susceptible to multiple interpretations and therefore truly ambiguous yet, given the context in which it was negotiated, not susceptible to a clearly contradicting and wholly unpersuasive interpretation asserted by the proponent of extrinsic evidence. In such a case, it seems clear that a court should exclude that evidence as violating the parol evidence rule despite the presence of some contract ambiguity. Finally, and most important, the ambiguity determination distracts the court from its primary objective — to enforce the contract as intended by the parties. Consequently, although relevant, contract ambiguity is not the only linchpin of a court’s decision to admit parol evidence.

The better rule is that the judge first considers the offered evidence and, if he or she finds that the contract language is “reasonably susceptible” to the interpretation asserted by its proponent, the evidence is admissible to determine the meaning intended by the parties. See Restatement § 215 cmt. b; see also Pacific Gas & Elec. Co. v. G. W. Thomas Dray. & Rigging Co., 69 Cal.2d 33, 69 Cal.Rptr. 561, 564, 566, 567-68, 442 P.2d 641, 644, 645-46 (1968);2 cf. Melson, 135 Ariz. at 121, 659 P.2d at 1266 (“A contract should be read in light of the parties’ intentions as reflected by their language and in view of all the circumstances.”). The meaning that appears plain and unambiguous on the first reading of a document may not appear nearly so plain once the judge considers the evidence. In such a case, the parol evidence rule is not violated because the evidence is not being offered to contradict or vary the meaning of the agreement. To the contrary, it is being offered to explain what the parties truly may have intended. We believe that this rule embodies the concepts endorsed by Corbin and adopted by this court ten years ago in Melson. Other courts more recently have expressed approval of the position taken by Corbin and the Restatement (Second) of Contracts. *155 See, e.g., C.R. Anthony Co. v. Loretto Mall Partners, 817 P.2d 238, 241-44 & n. 3 (N.M.1991); Isbrandtsen v. North Branch Corp., 150 Vt. 575, 556 A.2d 81, 83-85 (1988); Berg v. Hudesman, 115 Wash.2d 657, 801 P.2d 222, 227-30 (1990); see also 3 Corbin § 542, at 105-112 (Supp.1992) (citing cases).

A judge may not always be in a position to rule on a parol evidence objection at first blush, having not yet heard enough relevant evidence on the issue. If this occurs, the judge might, for example, admit the extrinsic evidence conditionally, reserve ruling on the issue until enough relevant evidence is presented, or, if the case is being tried to a jury, consider the evidence outside the jury’s presence. See, e.g., Ariz.R.Evid. 103(c), 104(b), 104(c), 105. Because the judge is in the best position to decide how to proceed, we leave this decision to his or her sound discretion. As noted also, the judge need not waste much time if the asserted interpretation is unreasonable or the offered evidence is not persuasive. A proffered interpretation that is highly improbable would necessarily require very convincing evidence. In such a case, the judge might quickly decide that the contract language is not reasonably susceptible to the asserted meaning, stop listening to evidence supporting it, and rule that its admission would violate the parol evidence rule. See 3 Corbin § 542, at 112; § 579, at 420.

We now apply these principles to the facts of this case.

B. Was the release so clear that the trial judge erred in admitting extrinsic evidence to interpret it?

Taylor released “all contractual rights, claims, and causes of action he ha[d] or may have against STATE FARM under the policy of insurance ... in connection with the collision ... and all subsequent matters.” See Appendix (emphasis added). Taylor argued that the bad faith claim sounds in tort and was therefore neither covered nor intended to be covered by the language releasing “all contractual” claims. The trial court found that

[pjarts of the document suggest that the parties contemplated the question of the insurer’s settlement of claims within policy limits. Yet on page 2 of the document, the release satisfies “all contractual rights, claims, and causes of action.” As a matter of statutory or contract construction, the word “contractual” modifies the words “rights,” “claims,” and the words “causes of action.” Although the breach of the duty of good faith and fair dealing arises out of contract, the action itself is a tort claim. Thus, there is ambiguity here. Where there is an ambiguity, parol evidence will be admitted on this issue.

Minute Entry, Jan. 12, 1987 at 1-2 (citations omitted). The court of appeals held that Taylor’s bad faith action was purely contractual and therefore, unlike the trial judge, found no ambiguity in the release language. Taylor, mem. dec. at 16.3 We must decide whether the release language is reasonably susceptible to Taylor’s proffered interpretation in light of the evidence relevant to the parties’ intent. If it is, admission of extrinsic evidence supporting his interpretation did not violate the parol evidence rule.

1. Was the release language reasonably susceptible to differing interpretations, including that the bad faith claim was not released despite the contractual quality of such a claim?

First, we address the court of appeals’ holding that Taylor’s bad faith claim was *156only contractual and that the trial court erred by finding that the release language, which indisputably covered contractual matters, was unclear, requiring extrinsic evidence for interpretation. Taylor, mem. dec. at 16, 23.

The court of appeals held that to assert a bad faith claim, Taylor first had to establish a “breach of contract”; that is, he must have been able to show that State Farm “denied, failed to pay, or failed to process a valid claim.” Id. at 16. The court held that at the time Taylor made the release agreement, his bad faith claim was contractual in nature. The release therefore unambiguously included bad faith, and parol evidence supporting the contrary was inadmissible. Id. at 23.4 We disagree. The proper inquiry is not whether the claim was contractual in nature or whether the judge believed it was contractual but, instead, what the parties intended to release when they used language conspicuously less inclusive than the release of “all claims.” Of course, the true doctrinal nature of the claim, if it could be determined, would be relevant evidence in the search for the parties’ contracting intent. If bad faith was ordinarily and universally thought of as a “contract” claim, it would be very difficult for Taylor to argue that the specific language in the agreement, releasing all “contractual” matters, was reasonably susceptible to his interpretation. See Producers Dairy Delivery Co. v. Sentry Ins. Co., 41 Cal.3d 903, 226 Cal.Rptr. 558, 563, 718 P.2d 920, 925 (1986) (noting parol evidence must be capable of persuading a reasonable person that the language meant something other than its ordinary meaning); General Motors Corp. v. Superior Court, 12 Cal.App. 4th 435, 15 Cal.Rptr.2d 622, 625-27 (Ct.App.1993) (applying plain language in a release because there was no competent evidence suggesting that the parties intended something different); Restatement § 212 cmt. b (noting that although the transaction is important, “the words of an integrated agreement remain the most important evidence of intention”). In such a case, Taylor would be in the unenviable position of arguing that “X” in fact does not mean “X.” This, however, is not the case.

It is true that bad faith has its genesis in contract. See Rawlings v. Apodaca, 151 Ariz. 149, 153-54, 726 P.2d 565, 569-70 (1986). Our cases show, however, that the precise legal character of a bad faith claim may depend on the context of the discussion. See, e.g., Deese v. State Farm Mut. Auto. Ins. Co., 172 Ariz. 504, 509, 838 P.2d 1265, 1270 (1992) (noting that a breach of an express covenant is not “a necessary prerequisite” for a bad faith claim); Sparks v. Republic Nat. Life Ins. Co., 132 Ariz. 529, 544, 647 P.2d 1127, 1142 (holding that a bad faith claim arises out of contract for attorney fee statute), cert. denied, 459 U.S. 1070, 103 S.Ct. 490, 74 L.Ed.2d 632 (1982); Noble v. National Am. Life Ins. Co., 128 Ariz. 188, 190, 624 P.2d 866, 868 (1981) (holding that a bad faith claim is a tort). Despite the contractual origin of an insurance bad faith claim, the seminal Arizona decision on the subject, less than six months old at the time State Farm and Taylor made the agreement, declared that such conduct is a tort. Noble, 128 Ariz. at 190, 624 P.2d at 868 (Feb. 17, 1981). Although Sparks, Rawlings, and Deese post date the execution of the release, these cases demonstrate the uninterrupted recognition that bad faith has components of both tort and contract and underscore the lack of consensus in the area. See Noble, 128 Ariz. at 191, 624 P.2d at 869 (Struckmeyer, C.J., dissenting) (“ ‘The relationship between the remedies in contract and tort presents a very confusing field, still in process of development, in which few courts have made any attempt to chart a path.’ ”) (quoting William L. Prosser, The *157Law of Torts § 92, at 614 (4th ed. 1971)). In fact, State Farm implicitly recognizes the dual nature of Taylor’s bad faith claim by conceding its tortious quality for purposes of its statute of limitations argument. See State Farm’s Opening Brief to Court of Appeals at 36-37 (alleging that Taylor’s bad faith claim is barred by the tort statute of limitations). State Farm also abandons the claim’s contractual side in arguing that attorney fees should not have been awarded. Id. at 49-50.

Because the legal character of bad faith was and is not universally established, the release reasonably could be interpreted as Taylor asserts. The trial court, therefore, did not err in concluding that the text of the release did not necessarily cover claims for bad faith.5

2. Was there extrinsic evidence to support the conclusion that the release language was reasonablg susceptible to Taglor’s interpretation?

At the time the parties made the agreement, it was obvious that Taylor had a sizable potential claim for bad faith. The release was agreed to months after the jury returned verdicts against Taylor and less than six months after this court recognized the tort of bad faith in Noble. Most of the alleged conduct that is the basis for Taylor’s bad faith claim had already occurred. In fact, although occurring later, the Rings garnished State Farm seeking to satisfy their entire judgment, including the excess above the policy limits, based on State Farm’s liability to Taylor for alleged bad faith. Ring v. State Farm Mut. Auto. Ins., 147 Ariz. 32, 33, 708 P.2d 457, 458 (Ct.App.1985). There was some evidence that Hofmann, on behalf of State Farm, directed that “general release language” be used in the agreement without expressly mentioning “bad faith.” 6 The document’s cryptic language supports this. Such a direction, in light of the obvious nature of the claim, supports Taylor’s interpretation.

Also, State Farm internally designated the $15,000 payment to Taylor as being for uninsured motorist (“UM”) coverage. Although State Farm denies its importance, its ultimate significance was for the fact finder to determine. State Farm’s subsequent conduct may shed light on its understanding of what was covered by the agreement. See Darner, 140 Ariz. at 393, 682 P.2d at 398. This, too, supports the idea that the release is reasonably susceptible to Taylor’s interpretation.

The potential size of the bad faith claim also cannot be ignored. At the time the parties entered into the release agreement, a jury had already rendered verdicts against Taylor far exceeding his insurance policy limits. Thus, the potential size of Taylor’s bad faith claim was obvious. We recognize that, with few exceptions, parties are free to structure a deal in any way they wish. Nevertheless, it is arguably reasonable to conclude that Taylor and his counsel would seek something more than just the payment of a potentially bona fide $15,-*158000 UM claim to release a bad faith claim possibly worth millions of dollars.

Finally, and perhaps most telling, is the fact that the parties used limiting language in the release. It is reasonable to believe that if the parties had agreed to release the bad faith claim, they would not have drawn the release so narrowly — confining it to “contractual” and “subsequent” matters, with no mention of tort claims or bad faith. Surely, State Farm knew what language would effectively release it from Taylor’s potential bad faith claim. It can be inferred that sophisticated parties in the business of settling insurance claims, faced with the task of releasing a claim as large as Taylor’s, would have used more specific or at least broader 7 language if that was their agreement. This is especially true in light of the conspicuous nature of the claim, the parties’ obvious knowledge of its existence, and the very recent supreme court authority characterizing the claim as a tort. Compare Edwards v. Comstock Ins. Co., 205 Cal.App.3d 1164, 252 Cal.Rptr. 807, 808, 810 (1988) (holding that broad language releasing “all” claims “whether in contract, tort or otherwise” necessarily included a bad faith claim) with Asare v. Hartford Fire Ins. Co., 1 Cal.App. 4th 856, 2 Cal.Rptr.2d 452, 455-56 (1991) (finding a release of workers’ compensation claim did not necessarily include the release of a discrimination claim, noting, among other things, that although aware of the latter, the release made “no explicit reference” to the discrimination claim). For these reasons, we hold that extrinsic evidence produces support for Taylor’s contention that the release language was not intended to release his bad faith claim.

This is not to say, however, that the release language excluded bad faith as a matter of law. Substantial evidence supports State Farm’s interpretation as well. The release language is broad enough to release something more than just the contractual UM claim.8 Also, the recitals refer to matters that may not pertain to UM coverage but are relevant to Taylor’s bad faith claim.9 Finally, there was credible evidence that both parties contemplated that bad faith would be covered by the release. All of the evidence, however, does not render the release language impervious to Taylor’s interpretation. Instead, it demonstrates that there were three reasonable, but conflicting, interpretations of the language used in the agreement: (1) the parties agreed to release the bad faith claim; (2) the parties agreed to exclude the bad faith claim; and (3) the parties did not reach any agreement regarding release of the bad faith claim (in which case, of course, the claim would not be released). In light of this, the trial judge correctly concluded that the release could not as a matter of law be interpreted to include or exclude Taylor’s bad faith claim.

C. Was the parol evidence for the jury?

Whether contract language is reasonably susceptible to more than one *159interpretation so that extrinsic evidence is admissible is a question of law for the court. See Leo Eisenberg & Co., Inc. v. Payson, 162 Ariz. 529, 532-33, 785 P.2d 49, 52-53 (1989). We have concluded in the preceding section that the language of the agreement, illuminated by the surrounding circumstances, indicates that either of the interpretations offered was reasonable. Because interpretation was needed and because the extrinsic evidence established controversy over what occurred and what inferences to draw from the events, the matter was properly submitted to the jury. See Burkons, 168 Ariz. at 351, 813 P.2d at 716; Leo Eisenberg, 162 Ariz. at 533, 785 P.2d at 53. The trial judge, therefore, instructed the jury as follows:

... [State Farm] has alleged the affirmative defense of release. In this regard, [State Farm] contends that the agreement ... was intended by the parties thereto to, among other things, release [State Farm] from all bad faith claims.

... Whether the parties intended the bad faith claims to be released is for you to determine. If you find that the parties to said agreement intended thereby that [State Farm] be released from bad faith claims, then your verdict must be for [State Farm].

A release is to be construed according to the intent of the parties to it. This intention is to be determined by what was within the contemplation of the parties when the release was executed, which, in turn, is to be resolved in the light of all of the surrounding facts and circumstances under which the parties acted.

Reporter’s Transcript, Mar. 12, 1987, at 184-85.10 The instruction states the issue quite clearly. So instructed, the jury resolved the release issue in Taylor’s favor.11 We leave that resolution undisturbed.

CONCLUSION

The trial court properly considered and then admitted extrinsic evidence to interpret the release and determine whether it included Taylor’s bad faith claim. That question, in this case, was appropriately left to the trier of fact. There remain other issues not resolved by the court of appeals. We are aware of the frustration that additional delay imposes on all involved, especially in a case as old as this. Nevertheless, because the other issues were initially and fully presented to the court of appeals, prudence dictates that the court of appeals complete its review of this case as expeditiously as possible. The decision of the court of appeals pertaining to the release is vacated and the matter is remanded to the court of appeals for resolution of the remaining issues.12

MOELLER, V.C.J., ZLAKET, J., and JAMES D. HATHAWAY, Judge, concur.

Justice FREDERICK J. MARTONE did not participate in this matter; pursuant to Ariz. Const, art. VI, § 3, Judge JAMES D. HATHAWAY of the Court of Appeals, Division Two, was designated to sit in his stead.

CORCORAN, Justice,

specially concurring:

I concur with the opinion — but without enthusiasm. It is certainly true that wavering and overlapping lines of interpretation, rather than bright straight borders prevail in this area of contract interpreta*160tion. I don’t know whether our opinion helps.

The canon of interpretation which we propound today is amorphous. The problem with an amorphous rule is that in the end, only this court can make a final determination in construing any contract. Our interpretation will be based upon which parol evidence impresses us the most. That ultimately means that this court must decide every contract dispute subject to this analysis. As the history of this case shows, the trial court may go one way and the court of appeals another and this court yet another.

I fear that this opinion makes this court the supreme court of arguments “that white is black and that a dollar is fifty cents” — to use the colorful words of Professor Corbin.

APPENDIX

Full Text of Release:

AGREEMENT

This Agreement made and entered into this 4tii day of August, 1981, by and between BOBBY SID TAYLOR and the STATE FARM MUTUAL AUTOMOBILE INSURANCE COMPANY, (hereinafter referred to as STATE FARM), by and through its agent undersigned.

WHEREAS, BOBBY SID TAYLOR was covered by an automobile insurance policy issued by STATE FARM, which was in effect on the 9th day of April, 1977, providing liability and uninsured motorist coverage to him, and

WHEREAS, an automobile collision occurred on April 9, 1977 between vehicles operated by BOBBY SID TAYLOR, DOUGLAS ALAN WISTROM and ANNE L. RING, and

WHEREAS, a trial took place in the Superior Court of Maricopa County, State of Arizona in consolidated causes C-382960 and G-383090, resulting in a jury verdict against BOBBY SID TAYLOR in the total amount of $2,621,000, and judgments having been entered against BOBBY SID TAYLOR in accordance with said jury verdicts, and

WHEREAS, having been fully apprised of all settlement offers made by the plaintiffs in the consolidated cases referred to above, during the discovery process, prior to trial, during the trial, and subsequently, BOBBY SID TAYLOR maintained and does now maintain that the operation of his motor vehicle on April 9, 1977 did not contribute to the injuries sustained by the plaintiffs, and at no time has he insisted, demanded, or even encouraged his insurer to settle the plaintiffs’ claims within his policy limits, and

WHEREAS, one of the drivers of an automobile involved in the collision on April 9, 1977, to wit: DOUGLAS ALAN WIS-TROM, was uninsured on the date of said collision, and BOBBY SID TAYLOR having a bona fide belief that the negligence of DOUGLAS ALAN WISTROM contributed to his bodily injuries sustained in that collision, and

WHEREAS, BOBBY SID TAYLOR has demanded compensation from STATE FARM under the uninsured motorist coverage afforded to him, and

WHEREAS, BOBBY SID TAYLOR desires to settle the uninsured motorist claim, and to relieve STATE FARM of any and all other contractual claims, interests, or causes of action he has or may have against STATE FARM, and

WHEREAS, STATE FARM has agreed that uninsured motorist coverage is available to BOBBY SID TAYLOR and appropriate under the facts surrounding the collision on April 9, 1977, and STATE FARM having been fully apprised in the premises,

THEREFORE, in consideration of the mutual [covenants contained herein, STATE FARM agrees to pay the sum of $15,000 to BOBBY SID TAYLOR in full satisfaction of all contractual rights, claims, and causes of action he has or may have against STATE FARM under the policy of insurance referred to herein, in connection with the collision on April 9, 1977, and all subsequent matters, and BOBBY SID TAYLOR hereby accepts that sum pursuant to the recitals contained herein.

[SIGNATURES]

CHIEF JUSTICE TURNAGE

delivered the Opinion of the Court.

This action arises out of a construction contract on which plaintiff Sherrodd, Inc., was a subcontractor. Sherrodd, Inc., appeals from a summary judgment entered for defendants by the District Court for the Thirteenth Judicial District, Yellowstone County. We affirm.

The issue is whether the entry of summary judgment for defendants was proper.

Sherrodd, Inc. (Sherrodd), is a family-owned Montana construction corporation. Sherrodd subcontracted with COP Construction (COP) to do certain earth-moving work involved in the construction of fifty family housing units in Forsyth, Montana, for the Army Corps of Engineers. COP itself was a subcontractor to the general contractors Morrison-Knudsen Company, Inc. (Morrison-Knudsen), and Schlekeway Construction, Inc. (Schlekeway). Safeco Insurance Company of America (Safeco) provided COP’s payment bond on the job.

Sherrodd contends that while its officer William Sherrodd was examining the building site in preparation for submitting a bid on this project, a representative of Morrison-Knudsen told him that there were 25,000 cubic yards of excavation to be performed on the job. It claims that its bid of $97,500 on the subcontract was made in reliance on that representation, based on $3.90 per cubic yard for 25,000 cubic yards. Morrison-Knudsen denies that its representative made any such statement to William Sherrodd.

*284Sherrodd’s bid, and, in turn, OOP’s bid including Sherrodd’s bid, were submitted and accepted. Sherrodd began work before a written contract was signed. While performing the earthwork, Sherrodd discovered that the quantity of work far exceeded 25,000 cubic yards.

The written contract between Sherrodd and COP provided that Sherrodd would perform earthwork in the quantity “LS” for the consideration of $97,500. The parties agree that the letters “LS” mean lump sum. Sherrodd contends that its officers signed the contract, even though by then they knew that the job involved more than 25,000 cubic yards of earthwork, because a COP officer threatened to withhold payment for work already done unless the contract was signed. Sherrodd further contends that the COP officer verbally represented that a deal would be worked out wherein Sherrodd would be paid more than the sum provided for in the contract. COP’s position is that it only agreed to assist Sherrodd in presenting a claim for additional compensation to the Army Corps of Engineers, based on differences in the moisture content of the soil from that stated in the bid proposal. That was done, but the claim was denied.

In its “Standard Subcontract Provisions,” the contract entered between Sherrodd and COP also provided that

’’the Subcontractor has, by examination, satisfied himself as to the ... character, quantity and kind of materials to be encountered ... No verbal agreement with any agent either before or after the execution of this Subcontract shall affect or modify any of the terms or obligations herein contained and this contract shall be conclusively considered as containing and expressing all of the terms and conditions agreed upon by the parties hereto. No changes ... shall be valid ... unless reduced to writing and signed by the parties hereto.”

Sherrodd was paid the $97,500 provided for in the contract, less approximately $9,750 for work left uncompleted. It brought this suit to set aside the price provisions in the contract and to recover quantum meruit plus tort damages. Its legal theories were fraud, both actual and constructive, and breach of the covenant of good faith and fair dealing. Defendants moved for summary judgment, which was granted based on the parol evidence rule regarding modification of written contracts.

Summary judgment is proper when the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there are no genuine issues of material fact and that the moving party is entitled to judgment as a matter of law. Rule 56(c), M.R.Civ.P. The District Court held that, under the *285parol evidence rule, Sherrodd could not introduce evidence of the alleged oral misrepresentations by either the Morrison-Knudsen representative or the COP officer. Therefore, it concluded that even taking the evidence in the light most favorable to Sherrodd, summary judgment for defendants was proper.

The parol evidence rule is codified in Montana statutes. Section 28-2-904, MCA, provides that:

“The execution of a contract in 'writing, whether the law requires it to be written or not, supersedes all the oral negotiations or stipulations concerning its matter which preceded or accompanied the execution of the instrument.”

Section 28-2-905, MCA, provides that when an agreement has been reduced to writing by the parties, there can be no evidence of the terms of the agreement other than the contents of the writing except when a mistake or imperfection of the writing is claimed or when the validity of the agreement is the fact in dispute.

Although it mentions mutual mistake in its brief to this Court, Sherrodd did not rely on that theory in the proceedings below, as evidenced in the pretrial order and in the District Court’s memorandum on the summary judgment. We will not consider on appeal a theory not raised at the trial court level. Morse v. Cremer (1982), 200 Mont. 71, 81, 647 P.2d 358, 363.

A further exception is made to the parol evidence rule when fraud is alleged. Section 28-2-905(2), MCA. However, that exception only applies when the alleged fraud does not relate directly to the subject of the contract. Where an alleged oral promise directly contradicts the terms of an express written contract, the parol evidence rule applies. Continental Oil Co. v. Bell (1933), 94 Mont. 123, 133, 21 P.2d 65, 67. Accord, Superior Oil Company v. Vanderhoof(D. Mont. 1969), 297 F.Supp. 1086.

Here, any reliance on the alleged fraudulent statement of the Morrison-Knudsen representative is contradicted by the terms of the written contract that Sherrodd has, “by examination, satisfied himself as to the ... character, quantity and kind of materials to be encountered.” The contention that the $97,500 covered only 25,000 cubic yards of earthwork contradicts the terms of the written agreement that all “negotiations and agreements” prior to the date of the contract are merged in the writing and that the work to be done is ‘lump sum.” We conclude that the parol evidence rule applies. Because the written agreement supersedes all previous oral *286agreements, the rule prohibits admission of any evidence of the representation by the Morrison-Knudsen representative.

Next we consider Sherrodd’s claim that COP officers induced Sherrodd officers to sign the contract with the promise that more money would be paid than the contract provided. Section 28-2-1602, MCA, provides that a written contract may be altered only by a subsequent contract in writing or by an executed oral agreement. Also, Sherrodd’s subcontract provided that “No changes ... shall be valid ... unless reduced to writing and signed by the parties hereto.” As the District Court noted, there is no allegation of a subsequent contract in writing, and if there had been an executed oral agreement to pay additional sums for the work, there would have been no reason for this lawsuit.

Because of the inadmissibility of Sherrodd’s evidence as to alleged misrepresentations, the claim of breach of the covenant of good faith and fair dealing also fails. There is no allegation of any violation of the express terms of the written contract, as would be required in this arms-length contract under our opinion in Story v. City of Bozeman (1990), 242 Mont. 436, 791 P.2d 767.

As we have stated,

“Commercial stability requires that parties to a contract may rely upon its express terms without worrying that the law will allow the other party to change the terms of the agreement at a later date.”

Baker v. Bailey (1989), 240 Mont. 139, 143, 782 P.2d 1286, 1288.

The parol evidence rule is the public policy of Montana and it is clearly established by statute and the decisions of this Court. If this public policy and rule is not upheld, contracting parties that include lawful provisions in written contracts would be under a cloud of uncertainty as to whether or not their written contracts may be relied upon. The public policy and law does not permit such uncertainty to occur.

We conclude that the compensation of Sherrodd is governed exclusively by the written contract and that Sherrodd’s claims are barred under the parol evidence rule. We hold that the District Court did not err in granting summary judgment for defendants.

Affirmed.

JUSTICES HARRISON, GRAY, McDONOUGH and WEBER concur.

JUSTICE TRIEWEILER,

dissenting:

I dissent from the opinion of the majority.

*287If the facts are as alleged by the plaintiff (and for purposes of this proceeding we must assume that they are), then the result of this case is that no party can be held accountable for its fraudulent conduct so long as it is in a sufficiently superior bargaining position to compel its victim to sign a document relieving it of liability.

The facts, as alleged by the plaintiff, offend any reasonable sense of fairness. No court should be so bound by a 58-year-old precedent that it cannot adapt to circumstances such as those presented in this case.

The plaintiff was informed by Lou Castino, the construction manager for Schlekeway and Associates, that the project he was being asked to bid on involved moving 25,000 cubic yards of dirt. It was based on that information that he submitted his bid. It was based on his bid that he was given an oral request to proceed with the work.

After commencing work on the project, plaintiff realized that the amount of earth that had to be moved greatly exceeded 25,000 cubic yards, and was actually more than twice that amount. He had conversations with representatives of both COP Construction and Schlekeway and Associates, during which it was agreed that the amount of work to be performed would be recalculated, and during which the defendants agreed to compensate plaintiff on the basis of the actual amount of work done, rather than the price which was originally agreed upon.

By May 22,1985, plaintiff had already been working on the project and had incurred substantial expenses and obligations to his own employees. He had not been paid for his work, and was still operating without a written agreement. It was on that date that he was requested by COP Construction’s superintendent to sign the written contract which the defendants now assert as a bar to his cause of action. He was advised that if he did not sign the agreement he would not receive the progress payment in the amount of $70,372.80 which was due. Without the progress payment he would not have been able to pay his current expenses and payroll.

He was further advised that he would not be bound by the terms of the written agreement, but that he would be paid for the actual work done at the rate of $3.90 per cubic yard.

Thereafter, the amount of earth work to be done was recalculated at approximately 50,000 cubic yards. On that basis, plaintiff tried to recover the full amount due, but payment was refused. Instead, the defendants raised the written agreement as a bar to any further payment to the plaintiff.

*288Because of the defendants’ failure to pay the plaintiff the additional $100,000 to $120,000 which they owed him, plaintiff’s business lost its ability to borrow money, lost its bonding, and was unable to complete additional contracts because of a lack of operating capital. Plaintiff was unable to bid on contracts that required bonding, and completely lost its ability to carry on business as it had in the past. As a direct result of the defendants’ failure to pay the amounts due, plaintiff was unable to continue in business as a construction company, which it had done for the previous 30 years.

If the plaintiff’s allegations are true, then defendant COP Construction Company’s conduct, at least, satisfies the elements of fraud. See Poulsen, et al. v. Treasure State Industries, 192 Mont. 69, 626 P.2d 822 (1981). COP’s employees represented to the plaintiff that he would be paid for the full amount of work done, regardless of the written terms of the contract. That representation was untrue and material, and COP’s superintendent either knew it was untrue or had no reason to believe that it was true. COP Construction intended that the plaintiff act in reliance upon that representation. Plaintiff did rely on it, and had no reason to believe that COP’s superintendent would mislead him. As a result, plaintiff has sustained the total loss of his business and substantial damages.

The majority has affirmed the dismissal of plaintiff’s claim based solely on the parol evidence rule found at § 28-2-904, MCA. That rule provides that a written agreement supersedes all oral negotiations which preceded or accompanied the execution of the instrument. Furthermore, § 28-2-905, MCA, provides that the terms of a written agreement cannot be proven by evidence other than what is contained in the written document.

However, an important exception is found at § 28-2-905(2), MCA, which provides, in relevant part, as follows:

“This section does not exclude other evidence of the circumstances under which the agreement was made or to which it relates... or other evidence to explain ... fraud.”

In addition, § 28-2-1611, MCA, provides as follows:

“When, through fraud or a mutual mistake of the parties or a mistake of one party while the other at the time knew or suspected, a written contract does not truly express the intention of the parties, it may be revised on the application of a party aggrieved so as to express that intention, so far as it can be done without prejudice to rights acquired by third persons in good faith and for value.”

(Emphasis added.)

*289In this case, in spite of the exceptions to the parol evidence rule set forth by statute above, the majority has chosen to rely on this Court’s 58-year-old decision in Continental Oil v. Bell, 94 Mont. 123, 133, 21 P.2d 65, 68 (1933). In that case, this Court held that parol evidence of fraud was not admissible when the oral promise directly contradicts a provision of the written contract.

I would not follow this Court’s previous decision in Continental Oil for two reasons:

1. That decision made no specific reference to the statute which is controlling, and yet adds qualifications to the statute which were not included by the legislature. The legislature provided that parol evidence could be offered to establish that a contract was induced by fraud. It made no exception where evidence of the fraudulent oral agreement contradicted a term in the written agreement.

2. To follow the decision in Continental Oil creates a terrible injustice, rewards fraudulent parties who are in a superior bargaining position, and totally defeats the purpose for which the fraud exception was provided to the parol evidence rule.

Based on this decision, and our previous decision in Continental Oil, all that a fraudulent party needs to do in order to avoid accountability for fraudulent conduct is to obtain the signature of his defrauded victim on a written agreement.

The majority expresses concern that but for this decision general contractors would not be able to rely on written agreements with their subcontractors. However, general contractors who induce subcontractors to enter into a written agreement by fraudulent representations should find no security in the piece of paper which resulted from their culpable conduct. Furthermore, a justice system worth its salt should have equal compassion for Montana’s many subcontractors who, while operating without the benefit of legal advice, sign whatever is necessary in order to keep their operations afloat and their crews at work. When what they have signed results from an obvious misrepresentation and causes them the kind of substantial damages and hardship that have resulted in this case, those subcontractors are entitled to the protection of Montana’s laws and its courts.

For these reasons, I dissent from the majority opinion. I would reverse the judgment of the District Court and remand for a jury trial to determine the merits of the plaintiff’s claim. That is really all the protection that Montana’s general contractors need.

JUSTICE HUNT concurs in the foregoing dissent.

HOFFMAN, District Judge:

Appellant Nanakuli Paving and Rock Company (Nanakuli) initially filed this breach of contract action against appellee Shell Oil Company (Shell) in Hawaiian State Court in February, 1976.1 Nanakuli, the second largest asphaltic paving contractor in Hawaii, had bought all its asphalt requirements from 1963 to 1974 from Shell under two long-term supply contracts; its suit charged Shell with breach of the later 1969 contract.2 The jury returned a verdict of $220,800 for Nanakuli on its first claim, which is that Shell breached the 1969 contract in January, 1974, by failing to price protect Nanakuli on 7200 tons of asphalt at the time Shell raised the price for asphalt from $44 to $76.3 Nanakuli’s theory is that price-protection, as a usage of the asphaltic paving trade in Hawaii, was incorporated into the 1969 agreement between the parties, as demonstrated by the routine use of price protection by suppliers to that trade, and reinforced by the way in which Shell actually performed the 1969 contract up until 1974. Price protection, appellant claims, required that Shell hold the price on the tonnage Nanakuli had already committed because Nanakuli had incorporated that price into bids put out to or contracts awarded by general contractors and government agencies. The District Judge set aside the verdict and granted Shell’s motion for judgment n. o. v., which decision we vacate. We reinstate the jury verdict because we find that, viewing the evidence as a whole, there was substantial evidence to support a finding by reasonable jurors that Shell breached its contract by failing to provide protection for Nanakuli in 1974. Quichocho v. Kelvinator Corp., 546 F.2d 812, 813 (9th Cir. 1976). We do not believe the evidence in this case was such that, giving Nanakuli the benefit of all inferences fairly supported by the evidence and without weighing the credibility of the witnesses, *778only one reasonable conclusion could have been reached by the jury. Cockrum v. Whitney, 479 F.2d 84, 85-86 (9th Cir. 1973).

Nanakuli offers two theories for why Shell’s failure to offer price protection in 1974 was a breach of the 1969 contract. First, it argues, all material suppliers to the asphaltic paving trade in Hawaii followed the trade usage of price protection and thus it should be assumed, under the U.C.C., that the parties intended to incorporate price protection into their 1969 agreement. This is so, Nanakuli continues, even though the written contract provided for price to be “Shell’s Posted Price at time of delivery,” F.O.B. Honolulu. Its proof of a usage that was incorporated into the contract is reinforced by evidence of the commercial context, which under the U.C.C. should form the background for viewing a particular contract. The full agreement must be examined in light of the close, almost symbiotic relations between Shell and Nanakuli on the island of Oahu, whereby the expansion of Shell on the island was intimately connected to the business growth of Nanakuli. The U.C.C. looks to the actual performance of a contract as the best indication of what the parties intended those terms to mean. Nanakuli points out that Shell had price protected it on the two occasions of price increases under the 1969 contract other than the 1974 increase. In 1970 and 1971 Shell extended the old price for four and three months, respectively, after an announced increase. This was done, in the words of Shell’s agent in Hawaii, in order to permit Nanakuli’s to “chew up” tonnage already committed at Shell’s old price.4

Nanakuli’s second theory for price protection is that Shell was obliged to price protect Nanakuli, even if price protection was not incorporated into their contract, because price protection was the commercially reasonable standard for fair dealing in the asphaltic paving trade in Hawaii in 1974. Observance of those standards is part of the good-faith requirement that the Code imposes on merchants in performing a sales contract. Shell was obliged to price protect Nanakuli in order to act in good faith, Nanakuli argues, because such a practice was universal in that trade in that locality.

Shell presents three arguments for upholding the judgment n. o. v. or, on cross appeal, urging that the District Judge erred in admitting certain evidence. First, it says, the District Court should not have denied Shell’s motion in limine to define trade, for purposes of trade usage evidence, as the sale and purchase of asphalt in Hawaii, rather than expanding the definition of trade to include other suppliers of materials to the asphaltic paving trade. Asphalt, its argument runs, was the subject matter of the disputed contract and the only product Shell supplied to the asphaltic paving trade.5 Shell protests that the judge, by expanding the definition of trade to include the other major suppliers to the asphaltic paving trade, allowed the admission of highly prejudicial evidence of routine price protection by all suppliers of aggregate.6 Asphaltic concrete paving is *779formed by mixing paving asphalt with crushed rock, or aggregate, in a “hot-mix” plant and then pouring the mixture onto the surface to be paved. Shell’s second complaint is that the two prior occasions on which it price protected Nanakuli, although representing the only other instances of price increases under the 1969 contract, constituted mere waivers of the contract’s price term, not a course of performance of the contract. A course of performance of the contract, in contrast to a waiver, demonstrates how the parties understand the terms of their agreement. Shell cites two U.C.C. Comments in support of that argument: (1) that, when the meaning of acts is ambiguous, the preference is for the waiver interpretation, and (2) that one act alone does not constitute a relevant course of performance. Shell’s final argument is that, even assuming its prior price protection constituted a course of performance and that the broad trade definition was correct and evidence of trade usages by aggregate suppliers was admissible, price protection could not be construed as reasonably consistent with the express price term in the contract, in which case the Code provides that the express term controls.

We hold that the judge did not abuse his discretion in defining the applicable trade, for purposes of trade usages, as the asphaltic paving trade in Hawaii, rather than the purchase and sale of asphalt alone, given the unusual, not to say unique, circumstances: the smallness of the marketplace on Oahu; the existence of only two suppliers on the island; the long and intimate connection between the two companies on Oahu, including the background of how the development of Shell’s asphalt sales on Oahu was inextricably linked to Nanakuli’s own expansion on the island; the knowledge of the aggregate business on the part of Shell’s Hawaiian representative, Bohner; his awareness of the economics of Nanakuli’s bid estimates, which included only two major materials, asphalt and aggregate; his familiarity with realities of the Hawaiian marketplace in which all government agencies refused to include escalation clauses in contract awards and thus pavers would face tremenduous losses on price increases if all their material suppliers did not routinely offer them price protection; and Shell’s determination to build Nanakuli up to compete for those lucrative government contracts with the largest paver on the island, Hawaiian Bitumuls (H.B.), which was supplied by the only other asphalt company on the islands, Chevron, and which was routinely price protected on materials. We base our holding on the reading of the Code Comments as defining trade more broadly than transaction and as binding parties not only to usages of their particular trade but also to usages of trade in general in a given locality. This latter seems an equitable application of usage evidence where the usage is almost universally practiced in a small market such as was Oahu in the 1960’s before Shell signed its 1969 contract with Nanakuli.7 Additionally, we hold that, under the facts of this case, a jury could reasonably have found that Shell’s acts on two occasions to price protect Nanakuli were not ambiguous and therefore indicated Shell’s understanding of the terms of the agreement with Nanakuli rather than being a waiver by Shell of those terms.8

*780Lastly we hold that, although the express price terms of Shell’s posted price of delivery may seem, at first glance, inconsistent with a trade usage of price protection at time of increases in price, a closer reading shows that the jury could have reasonably construed price protection as consistent with the express term. We reach this holding for several reasons. First, we are persuaded by a careful reading of the U.C.C., one of whose underlying purposes is to promote flexibility in the expansion of commercial practices and which rather drastically overhauls this particular area of the law. The Code would have us look beyond the printed pages of the contract to usages and the entire commercial context of the agreement in order to reach the “true understanding” of the parties. Second, decisions of other courts in similar situations have managed to reconcile such trade usages with seemingly contradictory express terms where the prior course of dealings between the parties, trade usages, and the actual performance of the contract by the parties showed a clear intent by the parties to incorporate those usages into the agreement or to give to the express term the particular meaning provided by those usages, even at times varying the apparent meaning of the express terms. Third, the delineation by thoughtful commentators of the degree of consistency demanded between express terms and usage is that a usage should be allowed to modify the apparent agreement, as seen in the written terms, as long as it does not totally negate it. We believe the usage here falls within the limits set forth by commentators and generally followed in the better reasoned decisions. The manner in which price protection was actually practiced in Hawaii was that it only came into play at times of price increases and only for work committed prior to those increases on non-escalating contracts. Thus, it formed an exception to, rather than a total negation of, the express price term of “Shell’s Posted Price at time of delivery.” Our decision is reinforced by the overwhelming nature of the evidence that price protection was routinely practiced by all suppliers in the small Oahu market of the asphaltic paving trade and therefore was known to Shell; that it was a realistic necessity to operate in that market and thus vital to Nanakuli’s ability to get large government contracts and to Shell’s continued business growth on Oahu; and that it therefore constituted an intended part of the agreement, as that term is broadly defined by the Code, between Shell and Nanakuli.

I.

History Of Nanakuli-Shell Relations Before 1973

Nanakuli, a division of Grace Brothers, Ltd., a Hawaiian corporation, is the smaller of the two major paving contractors on the island of Oahu, the larger of the two being Hawaiian Bitumuls (H.B.). Nanakuli first entered the paving business on Oahu in 1948, but it only began to move into the largest Oahu market, Honolulu, in the mid-1950’s. Until 1964 or so, Nanakuli only got small paving jobs, such as service stations, driveways, and small subdivision streets; it was not in a position to compete with H.B. for government contracts for major roads, airports, and other large jobs. In the early sixties Nanakuli owner Walter Grace began to negotiate a mutually advantageous arrangement with Shell whereby Shell, which had a small market percentage and no asphalt terminals in Hawaii,9 would sign a long-term supply contract with Nanakuli that would commit Nanakuli to buy its asphalt requirements from Shell. On the other hand, Nanakuli would be helped to expand its paving business on Oahu through a guaranteed supply and a discount on its asphalt prices. Nanakuli’s growth would *781expand the market for Shell’s asphalt on the island, which would justify Shell’s capital investment of a half a million dollars on Oahu, to which asphalt would be brought in heated tankers from Shell’s refinery in Martinez, California.10

Shell signed two five-year contracts in 1963: a supply contract with Nanakuli itself and a distributorship with Grace, which provided for a $2 commission on all Nanakuli’s sales. In fact, almost all Nanakuli’s sales were to itself and thus the commission operated, according to Shell’s Hawaiian representative, Bohner, primarily as a discount mechanism. Lennox, who succeeded Grace as president in 1965 at Grace’s death,11 testified that its purpose was “to make us competitive in our paving operation with our competitor [H.B.] who is much larger than ourselves because they were a distributor for the Standard Oil Company’s asphalt operation.” Lennox and Smith, who joined Nanakuli as vice-president in 1965 and eventually succeeded Lennox, both saw Nanakuli’s and Shell’s relationship as that of partners. That characterization was not denied by Bohner, Shell’s Hawaiian representative from 1964 to 1978, who, in fact, essentially corroborated their description of the close relations between the two companies. As a symbol of that relationship, Nanakuli painted its trucks “Shell white,” placed Shell’s logo on those trucks, chose the same orange as used by Shell for its own logo, and put the Shell logo on its stationary.

In 1966 Pacific Cement and Aggregates (P.C. & A.), Nanakuli’s landlord at its rented rock quarry at Halawa, was bought by Lone Star Cement Corporation, which later became Lone Star Industries. Lone Star requested that Nanakuli upgrade its plant facilities at the quarry, which Nanakuli estimated would cost between $250,000 to $300,000. Nanakuli, knowing Shell was eager to build up its paving business on Oahu, approached Shell for direct financing of the plant, an idea to which Shell was initially receptive. Lennox testified that Shell “had a sizeable installation at Iwilei and they didn’t think we were selling enough of their product and they wanted us to sell more and that’s why we enlarged our plant.” Shell management philosophy later changed and it was decided not to finance the plant directly but rather to offer Nanakuli an additional $2 volume discount on all sales over five thousand tons to help finance the plant, according to Bohner. Lennox testified that Shell authorized Nanakuli to tell Lone Star that Shell, with a million-dollar investment in Hawaii, fully supported Nanakuli’s plant expansion plans. In 1968 two top Shell asphalt officials came from the mainland to discuss Nanakuli’s expansion: Blee from San Francisco and Lewis from New York.12 Together with Bohner and Nanakuli’s Lennox and Smith, they met with officials of Nanakuli’s bank to discuss the loan and repayment schedule. The three contracts were finally signed after long negotiations on April 1, 1969. They were to parallel the amortization schedule of the bank loan for the plant: a supply contract, a distributorship contract, and volume discount letter, all three to last until December 31, 1975, at which point each would have the option to cancel on six-months’ notice, with a minimum duration of over seven years, April 1, 1969, to July 1, 1976. Such long-term contracts were certainly unusual for Shell and this one was *782probably unique among Shell’s customers, at least by 1974.13

Lennox’ testimony, which was partially stricken by the court as inadmissible, was that Shell’s agreement with Nanakuli in 1969 included a commitment by Shell never to charge Nanakuli more than Chevron charged H.B., in order to carry out the underlying purpose of the agreement to make Nanakuli competitive with H.B. and thus expand its and Shell’s respective businesses on Oahu. This testimony was ruled inadmissible as parol evidence.14 Shell, itself at the end of the trial, read in parts of Smith’s earlier deposition in which he made a similar point.15 Smith’s deposition testimony was that, although no written provision was included on price protection, he was led to believe by Shell’s Bohner that he would get price protection. “I think there was a thought running between the two parties at that time that something would *783be done.” It was only after the fact, he added, that Bohner told Nanakuli that he had to get permission for price protection from the mainland at the time of price increases; before that, “we had no idea how the pricing was done.” Smith’s comments on additional terms agreed to were not as probative as those of Lennox, who was president and chief negotiator for Nanakuli of the 1969 contract. Nanakuli’s offer of proof by Lennox was that Shell agreed “to sustain that price during those contracts in the same way in which Standard Oil Company [Chevron] sustained its price to the principal paving contractor on the island, Hawaiian Bitumuls, if there was any price increase” and to “not exceed Standard Oil’s [Chevron’s] posted price for asphalt sold to Hawaiian Bitumuls.” It was agreed that Nanakuli would have to submit bids on a fixed-price basis and be price protected to compete with H.B. “We understood that Shell gave us the same protection in our bidding and our purchase of asphalt from them to incorporate in our work,” Lennox did testify, adding that by that means Nanakuli could compete with H.B. and Shell would thereby benefit. He said Nanakuli understood the price term to mean that Shell would not increase prices without advance notice and would hold the price on work bid for enough time to allow Nanakuli to use up the tonnage bid at the old price. Smith’s testimony backed up that of Lennox: the price was to be “posted price as bid as was understood between the parties,” further explaining that it was to be Shell’s price at time and place of delivery, except for price increases, at which point the price was time and place of bid for a period of time or a specified tonnage.16

Much information relevant to the “commercial context” of the agreement, essential to an understanding of the meaning of the terms, was contained in the testimony of Smith and Lennox. See Haw.Rev.Stat. § 490:2-202, Comment 2. None of it was flatly contradicted by the evasive responses of Bohner to such inquiries. Smith testified that Bohner was in Nanakuli’s offices at least weekly, sometimes was involved in the preparation of bids, and knew when a bid had been submitted or a contract awarded, at times attending the awards himself. The amount of time he spent following Nanakuli’s progress was understandable; as he explained, in the 1960’s and 1970’s Nanakuli “was basically . . . our only customer at this time.” Thus Nanakuli felt it had no need to notify Shell in writing of projects awarded; Lennox testified he was only told to do so on one occasion, the award of its first big contract on August 17, 1970. Due to the long shipping time from California, Bohner had to know well in advance Nanakuli’s supply needs. Lennox testified that Bohner was “as close to us as any person could be as far as the asphalt supply was concerned. He had to know what we were doing . . . . ” Lennox said Bohner was not only aware of Nanakuli’s day-to-day progress but knowledgeable about the broader asphaltic paving marketplace in which Nanakuli was competing. Bohner denied having closely delved into the economics of the aggregate trade or even knowing much about the asphalt prices of its chief competitor, Chevron, but he admitted he was interested in Chevron’s prices; made inquiries of Chevron’s customers, in-*784eluding H.B.; and had a ballpark estímate of aggregate prices. The jury could have inferred that he knew more about Chevron’s and the aggregate suppliers’ prices than he admitted. The jury could have looked to the importance of aggregate prices to Nanakuli’s success in getting large contracts; Bohner’s statements that he was “very definitely interested in the construction trades,” especially asphaltic paving; his subscription to a local construction journal that listed paving as well as other construction projects on Oahu; his membership in an asphaltic paving institute; and his knowledge about the use of aggregates as a component of asphaltic paving.

II

Trade Usage Before And After 1969

The key to price protection being so prevalent in 1969 that both parties would intend to incorporate it into their contract is found in one reality of the Oahu asphaltic paving market: the largest paving contracts were let by government agencies and none of the three levels of government — local, state, or federal — allowed escalation clauses for paving materials. If a paver bid at one price and another went into effect before the award was made, the paving company would lose a great deal of money, since it could not pass on increases to any government agency or to most general contractors. Extensive evidence was presented that, as a consequence, aggregate suppliers routinely price protected paving contractors in the 1960’s and 1970’s, as did the largest asphaltic supplier in Oahu, Chevron. Nanakuli presented documentary evidence of routine price protection by aggregate suppliers as well as two witnesses: Grosjean, Vice-President for Marketing óf Ameron H.C. & D., and Nihei, Division Manager of Lone Star Industries for Pacific Cement and Aggregate (P.C. & A.). Both testified that price protection to their knowledge had always been practiced: at H.C. & D. for many years prior to Grosjean’s arrival in 1962 and at P.C.&A. routinely since Nihei’s arrival in 1960. Such protection consisted of advance notices of increases, coupled with charging the old price for work committed at that price or for enough time to order the tonnage committed. The smallness of the Oahu market led to complete trust among suppliers and pavers. H.C. & D. did not demand that Nanakuli or other pavers issue purchase orders or sign contracts for aggregate before incorporating its aggregate prices into bids. Nanakuli would merely give H.C. & D. a list of projects it had bid at the time H.C. & D. raised its prices, without documentation. “Their word and letter is good enough for us,” Grosjean testified. Nihei said P.C. & A. at the time of price increases would get a list of either particular projects bid by a paver or simply total tonnage bid at the old price. “We take either one. We take their word for it.” None of the aggregate companies had a contract with Nanakuli expressly stating price protection would be given; Nanakuli’s contract with P.C. & A. merely set out that P.C. & A. would not charge Nanakuli more than it charged its other customers.

The evidence about Chevron’s practice of price protection came in the form of an affidavit by Bery Jameyson, Chevron’s Division Manager-Asphalt in California. He stated that Chevron had routinely price protected H.B. on work bid for many years, the last occasion prior to the signing of the 1969 contracts between Nanakuli and Shell being a price increase put into effect on March 7, 1969, with the understanding that H.B. . would be protected on work bid, which amounted to 12,000 tons. In answer to Shell’s protest that such evidence was not relevant without the contract itself, Nanakuli introduced the contract into evidence. Much like the contract at issue here, it provided that the price to H.B. would be a given percentage of the price Chevron set for a specified crude oil in California. No mention was made of price protection in the written contract between H.B. and Chevron.

In addition to evidence of trade usages existing in 1969 when the contract at issue was signed, the District Judge let in evidence of the continuation of that trade usage after 1969, over Shell’s protest. He *785stated that, giving a liberal reading to Section 1 — 205, he felt that later evidence was relevant to show that the expectation of the parties that a given usage would be observed was justified. The basis for incorporating a trade usage into a contract under the U.C.C. is the justifiable expectation of the parties that it will be observed. That later evidence consisted here of more price protection by the aggregate companies on Oahu, as well as continued asphalt price protection. Chevron after 1969 continued price protecting H.B. on Oahu and, on raising prices in 1979, price protected Nanakuli on the island of Molokai, where Nanakuli purchased its asphalt from Chevron. Additionally, Shell price protected Nanakuli in 1977 and 1978 on Oahu.17

Ill

Shell’s Course Of Performance Of The 1969 Contract

The Code considers actual performance of a contract as the most relevant evidence of how the parties interpreted the terms of that contract. In 1970 and 1971, the only points at which Shell raised prices between 1969 and 1974, it price protected Nanakuli by holding its old price for four and three months, respectively, after announcing a price increase. In the late summer of 1970, Shell had announced a price increase from $35 to $40 a ton effective September 1, 1970. When Nanakuli protested to Bohner that it should be price protected on work already committed, Blee wrote Bohner an in-house memo that, if Bohner could not “convince” Nanakuli to go along with the price increase on September 1, he should try to “bargain” to get Nanakuli to accept the price raise by at least the first of the year, which was what was finally agreed upon. During that four-month period, Nanakuli bought 3,300 tons. Shell announced a second increase in October, 1970, from $40 to $42 effective December 31st. Before that increase went into effect, on November 25 Shell increased the raise to $4, making the price $44 as of the first of the year.18 Shell again agreed to price protect Nanakuli by holding the price at $40, which had been the official price since September 1, for three months from January to March, 1971. Shell did not actually raise prices again until January, 1974, but at several points it believed that increases would be necessary and gave several months’ advance notice of those possible increases. Those actions were in accord with Shell’s own policy, as professed by Bohner, and that of other asphalt and aggregate suppliers: to give at least several months’ advance notice of price increases. On January 14,1971, Shell wrote its asphalt customers that the maximum 1971 increase would be to $46. On July 9, 1971, another letter promised the price would not go over $50 in 1972. In addition, Bohner volunteered on direct the information that Shell price protected Nanakuli on the only two occasions of price increases after 1974 by giving 6 months’ advance notice in 1977 and 3 or 4 months’ advance notice in 1978, a practice he described as “in effect carryover pricing,” his term for price protection. By its actions, Bohner testified, Shell allowed Nanakuli time to make arrangements to buy up tonnage committed at the old price, that is, to “chew up” tonnage bid or contracted. Shell apparently offered this testimony to impress the jury with its subsequent good faith toward Nanakuli. In fact, it also may have reinforced the impression of the universality of price protection in the asphaltic paving trade on Oahu and, by *786showing Shell’s adherence to that practice on every relevant occasion except 1974, have highlighted for the jury what was the commercially reasonable standard of fair dealing in effect on Oahu in 1974.

IV

Shell-Nanakuli Relations, 1973-74

Two important factors form the backdrop for the 1974 failure by Shell to price protect Nanakuli: the Arab oil embargo and a complete change of command and policy in Shell’s asphalt management. The jury was read a page or so from the World Book about the events and effect of the partial oil embargo, which shortened supplies and increased the price of petroleum, of which asphalt is a byproduct. The federal government imposed direct price controls on petroleum, but not on asphalt. Despite the international importance of those events, the jury may have viewed the second factor as of more direct significance to this case. The structural changes at Shell offered a possible explanation for why Shell in 1974 acted out of step with, not only the trade usage and commercially reasonable practices of all suppliers to the asphaltic paving trade on Oahu, but also with its previous agreement with, or at least treatment of, Nanakuli.

Bohner testified to a big organizational change at Shell in 1973 when asphalt sales were moved from the construction sales to the commercial sales department. In addition, by 1973 the- top echelon of Shell’s asphalt sales had retired. Lewis and Blee, who had negotiated the 1969 contract with Nanakuli, were both gone.19 Their duties were taken over by three men: Fuller in San Mateo, California, District Manager for Shell Sales, Lawson, and Chippendale, who was Shell’s regional asphalt manager in Houston. When the philosophy toward asphalt pricing changed, apparently no one was left who was knowledgeable about the peculiarities of the Hawaiian market or about Shell’s long-time relations with Nanakuli or its 1969 agreement, beyond the printed contract.

Shell had begun rethinking its asphalt pricing policies several years before. Swanson, who succeeded Lewis in New York in 1970, wrote an internal memorandum on April 21,1970, in which he discussed frankly the advantages and disadvantages of price protection of its asphalt buyers. Such a practice assured Shell of captive-volume sales, he wrote. The practice of granting carry-over pricing at times of price increases, however, had the unfortunate side effect of depressing prices in the asphalt market everywhere else, the memorandum concluded. This rethinking apparently led to a November 25, 1970, letter setting out “Shell’s New Pricing Policy” at its Honolulu and Hilo terminals. The letter explained the elimination of price protection: “In other words, we will no longer guarantee asphalt prices for the duration of any particular construction projects or for the specific lengths of time. We will, of course, honor any existing prices which have been committed for specific projects for which we have firm contractual commitments.” The letter requested a supply contract be signed with Shell within 15 days of the receipt of an award by a customer.

The District Judge based his grant of judgment n. o. v.20 largely on his belief that, had Nanakuli desired price protection, it should have complied with Shell’s request in that 1970 letter, by which we assume he meant Nanakuli should have made a firm contractual commitment with Shell for each *787project on which its bid was successful within 15 days of award.21 That conclusion, however, ignores several facts. First, compliance by Nanakuli with the letter’s demand that a contract be signed within 15 days of an award would have offered Nanakuli little, if any, protection. Nanakuli still would have been stuck with only charging the government the price incorporated into its bid if Shell raised its price between bid and award. The purpose of price protection was to guarantee the price in effect when a paver made a bid because of the often lengthy time span between bid and award. Second, if price protection was a part of Nanakuli’s 1969 agreement with Shell, Shell had no right to terminate unilaterally that protection. Third, the letter was addressed to “Gentlemen” with Nanakuli’s name typed in at the top; it was apparently addressed to all Shell’s Hawaiian customers. Fourth, Nanakuli officials testified that they did not believe the letter was applicable to its unusual situation of already having a long-term contract with Shell. Smith and Lennox both testified that they did not view the letter as applicable to that supply contract but only to sales it might make to third parties under the distributorship contract. Shell characterized that argument as disingenuous, given Nanakuli’s infrequent, if not nonexistent, sales to third parties. Nevertheless, the letter does assume that a Shell customer would need to sign a contract or purchase order setting forth the terms of sale as well as the price for any asphalt they would need to buy after an award. Nanakuli, on the other hand, already had a supply contract with all the terms of sales set forth, a point its two officials made repeatedly at trial. For example, Smith testified he saw no need to notify Shell because Bohner knew of each project and because the supply contract was a firm contractual commitment with Shell.22 Shell had added in the 1970 letter; “All previous contractual commitments made prior to the date of this letter will, of course, be honored.” Smith’s reading of this was that Nanakuli’s supply contract with Shell was a firm contractual commitment by Shell and that no further contract was needed. “We felt that this letter was unapplicable [sic] to our supply contract, that we already had a contractual commitment with Shell Oil Company which was not to end before 1975.” Smith said he did not discuss with Bohner that part of the letter, which also announced the increase in asphalt prices to $44 on January 1, because “[t]here was no need to. The price had been protected before. He knew it. We knew it.” There is an additional reason why Nanakuli might have felt that the letter did not apply to its particular situation. The letter announced that Shell would charge “from this date forward . . . the posted selling price on the date of purchase.” This was different from the express term of Nanakuli’s contract with Shell, which was the price in effect at time of delivery. Either Shell’s agreement with Nanakuli embraced price protection, in which case Shell could not unilaterally abrogate Nanakuli’s rights by this letter, or, as Shell argues, only the words on the written contract counted, in which case the price to Nanakuli was Shell’s price at delivery. In the latter case, Nanakuli could safely ignore any attempt by Shell to change the price to that at purchase. Given Nanakuli’s particular agreement, as it understood the agreement to be, the jury could have believed that Nanakuli officials reacted reasonably in believing that parts of the letter dealing with the need to notify Shell of awards won did not apply to Nanakuli.

Nanakuli’s strongest argument as to its failure to comply with the letter was that there was no need to notify Shell, as Bohner *788already knew of each project as it was bid and each award as it was made. Lennox testified, “The Shell Oil representative was in our office frequently and knew what jobs we had successfully bid.” At another point Lennox said, “The Shell representative was in the office and was fully aware of what we were doing and what jobs we had gotten. He was familiar and was more or less a partner in this thing; he even attended the bid openings at times. He was fully aware and congratulated us every time we got a nice big job because it was more for Shell.” Bohner kept his principals informed of Nanakuli’s projects, Lennox said. He added, “[W]e had always been protected and our understanding was that we were protected and it wasn’t necessary to keep making notices.” Smith in his deposition said that Bohner only told him that he lacked the authority to grant price protection “after the fact.” Since he knew nothing about how Shell arrived at its pricing, Smith assumed Bohner could carry out Shell’s agreement to price protect Nanakuli each time it was needed without consulting the mainland.

After Shell’s December 31, 1973, letter arrived on January 4, 1974, Smith called Bohner, as he had done before at times of price increases, to ask for price protection, this time on 7200 tons.. Bohner told Smith that he would have to get in touch with the mainland, but he expected that the response would be negative. Smith wrote several letters in January and February asking for price protection. After getting no satisfaction, he finally flew to California to meet with Lawson, Fuller, and Chippendale. Chippendale, from the Houston office, was acknowledged by the other two to be the only person with authority to grant price protection.23 All three Shell officials lacked any understanding of Nanakuli and Shell’s long, unique relationship or of the asphaltic trade in Oahu. They had never even seen Shell’s contracts with Nanakuli before the meeting. When apprised of the three and their seven-year duration, Fuller remarked on the unusual nature of Nanakuli’s relations with Shell, at least within his district. Chippendale felt it was probably unique for Shell anywhere. Smith testified that Fuller admitted to knowing nothing, beyond the printed page of Nanakuli’s agreement with Shell, of the background negotiation or *789Shell’s past pricing policies toward Nanakuli. Chippendale could not understand why Nanakuli even had a distributorship contract giving it a $2 commission on sales; he thought Nanakuli had been paid “illegally.” No one had ever heard about Shell giving price protection to Nanakuli before. Instead of asking Bohner directly, Chippendale told Fuller to search the files for something on paper. Fuller testified that Shell would not act without written proof of Shell’s past price protection of Nanakuli. He admitted he was unable to find anything in the files before 1972 because the departments had been reorganized in that year, about which he informed Chippendale. Chippendale accordingly decided to deny Nanakuli any price protection and wrote a draft of a letter for Fuller to send Nanakuli. He wrote a note to Fuller that he should adopt the “least said” approach with Nanakuli and check any letters with the legal department. When asked at trial if he had ever simply asked Bohner about Shell’s past pricing practices toward Nanakuli, Fuller answered, “No, I didn’t know we had it, other than the standard policy if we had one which we didn’t.”24 Chippendale told Smith in the California meeting that, although 7200 tons represented an infinitesimal amount for Shell, it would set a bad precedent for Shell, since price protection was not Shell’s “current policy.” (emphasis supplied). Shell people told him, Smith testified from contemporaneously made notes, that “any past practice was inapplicable at the present time.”25 Smith testified from those same notes that he had left the meeting under the impression that Shell was going out of business in Hawaii.

We conclude that the decision to deny Nanakuli price protection was made by new Houston management without a full understanding of Shell’s 1969 agreement with Nanakuli or any knowledge of its past pricing practices toward Nanakuli. If Shell did commit itself in 1969 to price protect Nanakuli, the Shell officials who made the decisions affecting Nanakuli in 1974 knew nothing about that commitment. Nor did they make any effective effort to find out. They acted instead solely in reliance on the 1969 contract’s express price term, devoid of the commercial context that the Code says is necessary to aii understanding of the meaning of the written word. Whatever the legal enforceability of Nanakuli’s right, Nanakuli officials seem to have acted in good faith reliance on its right, as they understood it, to price protection and rightfully felt betrayed by Shell’s failure to act with any understanding of its past practices toward Nanakuli.

V

Scope Of Trade Usage

The validity of the jury verdict in this case depends on four legal questions. First, how broad was the trade to whose usages Shell was bound under its 1969 agreement *790with Nanakuli: did it extend to the Hawaiian asphaltic paving trade or was it limited merely to the purchase and sale of asphalt, which would only include evidence of practices by Shell and Chevron? Second, were the two instances of price protection of Nanakuli by Shell in 1970 and 1971 waivers of the 1969 contract as a matter of law or was the jury entitled to find that they constituted a course of performance of the contract? Third, could the jury have construed an express contract term of Shell’s posted price at delivery as reasonably consistent with a trade usage and Shell’s course of performance of the 1969 contract of price protection, which consisted of charging the old price at times of price increases, either for a period of time or for specific tonnage committed at a fixed price in non-escalating contracts? Fourth, could the jury have found that good faith obliged Shell to at least give advance notice of a $32 increase in 1974, that is, could they have found that the commercially reasonable standards of fair dealing in the trade in Hawaii in 1974 were to give some form of price protection?

We approach the first issue in this case mindful that an underlying purpose of the U.C.C. as enacted in Hawaii is to allow for liberal interpretation of commercial usages. The Code provides, “This chapter shall be liberally construed and applied to promote its underlying purposes and policies.” Haw.Rev.Stat. § 490:1-102(1). Only three purposes are listed, one of which is “[t]o permit the continued expansion of commercial practices through custom, usage and agreement of the parties; . .. . ” Id. § 490:l-102(2)(b). The drafters of the Code explain:

This Act is drawn to provide flexibility so that, since it is intended to be a semi-permanent piece of legislation, it will provide its own machinery for expansion of commercial practices. It is intended to make it possible for the law embodied in this Act to be developed by the courts in the light of unforeseen and new circumstances and practices. . . .

.... The text of each section should be read in the light of the purpose and policy of the rule or principle in question, as also of the Act as a whole, and the application of the language should be construed narrowly or broadly, as the case may be, in conformity with the purposes and policies involved.

.... [t]he Code seeks to avoid . . . interference with evolutionary growth

This principle of freedom of contract is subject to specific exceptions found elsewhere in the Act. . . . [An example being the bar on contractual exclusion of the requirement of good faith, although the parties can set out standards for same.] ... In this connection, Section 1-205 incorporating into the agreement prior course of dealing and usages of trade is of particular importance.

Id., Comments 1 & 2 (emphasis supplied). We read that to mean that courts should not stand in the way of new commercial practices and usages by insisting on maintaining the narrow and inflexible old rules of interpretation. We seek the. definition of trade usage not only in the express language of the Code but also in its underlying purposes, defining it liberally to fit the facts of the particular commercial context here.26

*791The Code defines usage of trade as “any practice or method of dealing having such regularity of observance in a place, vocation or trade as to justify an expectation that it will be observed with respect to the transaction in question.” Id. § 490:1-205(2) (emphasis supplied). We understand the use of the word “or” to mean that parties can be bound by a usage common to the place they are in business, even if it is not the usage of their particular vocation or trade. That reading is borne out by the repetition of the disjunctive “or” in subsection 3, which provides that usages “in the vocation or trade in which they are engaged or of which they are or should be aware give particular meaning to and supplement or qualify terms of an agreement.” Id. § 490:1 — 205(3). The drafters’ Comments say that trade usage is to be used to reach the “. . .. commercial meaning of the agreement. ...” by interpreting the language “as meaning what it may fairly be expected to mean to parties involved in the particular transaction in a given locality or in a given vocation or trade.” Id., Comment 4 (emphasis supplied). The inference of the two subsections and the Comment, read together, is that a usage need not necessarily be one practiced by members of the party’s own trade or vocation to be binding if it is so commonly practiced in a locality that a party should be aware of it. Subsection 5 also shows the importance of the place where the usage is practiced: “An applicable usage of trade in the place where any part of performance is to occur shall be used in interpreting the agreement as to that part of the performance.” The validity of this interpretation is additionally demonstrated by the Comment of the drafters: “Subsection (3), giving the prescribed effect to usages of which the parties ‘are or should be aware’, reinforces the provision of subsection (2) requiring not universality but only the described ‘regularity of observance’ of the practice or method. This subsection also reinforces the point of subsection (2) that such usages may be either general to trade or particular to a special branch of trade.” Id., Comment 7 (emphasis supplied). This language indicates that Shell would be bound not only by usages of sellers of asphalt but by more general usages on Oahu, as long as those usages were so regular in their observance that Shell should have been aware of them. This reading of the Code, in our opinion, achieves an equitable result. A party is always held to conduct generally observed by members of his chosen trade because the other party is justified in so assuming unless he indicates otherwise. He is held to more general business practices to the extent of his actual knowledge of those practices or to the degree his ignorance of those practices is not excusable: they were so generally practiced he should have been aware of them.

No U.C.C. cases have been found on this point, but the court’s reading of the Code language is similar to that of two of the best-known commentators on the U.C.C.:

Under pre-Code law, a trade usage was not operative against a party who was not a member of the trade unless he actually knew of it or the other party could reasonably believe he knew of it.

J. White & R. Summers, Uniform Commercial Code, § 12-6 at 371 (1972) (emphasis supplied) (citing 3 A. Corbin, Corbin on Contracts § 557 at 248 (I960)). See also Restatement of Contracts § 247, Comment b (1932); 5 S. Williston, Williston on Contracts § 661 at 113-18 (3d. ed. 1961). White and Summers add (emphasis supplied):

This view has been carried forward by 1-205(3), .... [U]sage of the trade is *792only binding on members of the trade involved or persons who know or should know about it. Persons who should be aware of the trade usage doubtless include those who regularly deal with members of the relevant trade, and also members of a second trade that commonly deals with members of a relevant trade (for example, farmers should know something of seed selling).

White & Summers, supra, § 12-6 at 371. Using that analogy, even if Shell did not “regularly deal” with aggregate supplies, it did deal constantly and almost exclusively on Oahu with one asphalt paver. It therefore should have been aware of the usage of Nanakuli and other asphaltic pavers to bid at fixed prices and therefore receive price protection from their materials suppliers due to the refusal by government agencies to accept escalation clauses. Therefore, we do not find the lower court abused its discretion or misread the Code as applied to the peculiar facts of this case in ruling that the applicable trade was the asphaltic paving trade in Hawaii. An asphalt seller should be held to the usages of trade in general as well as those of asphalt sellers and common usages of those to whom they sell. Certainly, under the unusual facts of this ease it was not unreasonable for the judge to extend trade usages to include practices of other material suppliers toward Shell’s primary and perhaps only customer on Oahu. He did exclude, on Shell’s motion in limine, evidence of cement suppliers. He only held Shell to routine practices in Hawaii by the suppliers of the two major ingredients of asphaltic paving, that is, asphalt and aggregate. Those usages were only practiced towards two major pavers. It was not unreasonable to expect Shell to be knowledgeable about so small a market. In so ruling, the judge undoubtedly took into account Shell’s half-million dollar investment in Oahu strictly because of a long-term commitment by Nanakuli, its actions as partner in promoting Nanakuli’s expansion on Oahu, and the fact that its sales on Oahu were almost exclusively to Nanakuli for use in asphaltic paving. The wisdom of the pre-trial ruling was demonstrated by evidence at trial that Shell’s agent in Hawaii stayed in close contact with Nanakuli and was knowledgeable about both the asphaltic paving market in general and Nanakuli’s bidding procedures and economics in particular.

Shell argued not only that the definition of trade was too broad, but also that the practice itself was not sufficiently regular to reach the level of a usage and that Nanakuli failed to show with enough precision how the usage was carried out in order for a jury to calculate damages. The extent of a usage is ultimately a jury question. The Code provides, “The existence and scope of such a usage are to be proved as facts.” Haw.Rev.Stat. § 490:l-205(2).27 The practice must have “such regularity of observance ... as to justify an expectation that it will be observed. . . . ” Id. The Comment explains:

The ancient English tests for “custom” are abandoned in this connection. Therefore, it is not' required that a usage of trade be “ancient or immemorial,” “universal” or the like .... [Fjull recognition is thus available for new usages and for usages currently observed by the great majority of decent dealers, even though dissidents ready to cut corners do not agree.

Id., Comment 5. The Comment’s demand that “not universality but only the described ‘regularity of observance’ ” is required reinforces the provision only giving “effect to usages of which the parties ‘are or should be aware’ . . . . ” Id., Comment 7. A “regularly observed” practice of protection, of which Shell “should have been aware,” was enough to constitute a usage that Nanakuli had reason to believe was incorporated into the agreement.28

*793Nanakuli went beyond proof of a regular observance. It proved and offered to prove 29 that price protection was probably a universal practice by suppliers to the asphaltic paving trade in 1969.30 It had been practiced by H.C. & D. since at least 1962, by P.C. & A. since well before 1960, and by Chevron routinely for years, with the last specific instance before the contract being March, 1969, as shown by documentary evidence. The only usage evidence missing was the behavior by Shell, the only other asphalt supplier in Hawaii, prior to 1969. That was because its only major customer was Nanakuli and the judge ruled prior course of dealings between Shell and Nanakuli inadmissible. Shell did not point in rebuttal to one instance of failure to price protect by any supplier to an asphalt paver in Hawaii before its own 1974 refusal to price protect Nanakuli. Thus, there clearly was enough proof for a jury to find that the practice of price protection in the asphaltic paving trade existed in Hawaii in 1969 and was regular enough in its observance to rise to the level of a usage that would be binding on Nanakuli and Shell.

Shell next argues that, even if such a usage existed, its outlines were not precise enough to determine whether Shell would have extended the old price for Nanakuli for several months or would have charged the old price on the volume of tonnage committed at that price. The jury awarded Nanakuli damages based on the specific tonnage committed before the price increase of 1974. Shell says the jury could not have ascertained with enough certainty how price protection was carried out to calculate such an award for Nanakuli. The argument is not persuasive. The Code provides, “The remedies provided by this chapter shall be liberally administered to the end that the aggrieved party may be put in as good a position as if the other party had fully performed. ...” Id. % 490:1-106(1). The Comments list as one of three purposes of this section “to reject any doctrine that damages must be calculable with mathematical accuracy. Compensatory damages are often at best approximate: they have to be proved with whatever definiteness and accuracy the facts permit, but no more.” Id., Comment 1. Nanakuli got advance notices of each but the disputed increase by Shell, as well as an extension of several months at the old price in 1970, 1971, 1977, and 1978. Shell protests that in 1970 and 1971 Nanakuli’s protected tonnage only amounted to 3,300 and 1,100 tons, respectively. Chevron’s price protection of H.B. in 1969 however, is also part of the trade usage; H.B.’s protection amounted to 12,-000 tons. The increase in Nanakuli’s tonnage by 1974 is explained by its growth since the 1970 and 1971 increases.

In addition, the scope of protection offered by a particular usage is left to the jury:

In cases of a well established line of usage varying from the general rules of this Act where the precise amount of the variation has not been worked out into a single standard, the party relying on the usage is entitled, in any event, to the minimum variation demonstrated. The whole is not to be disregarded because no particular line of detail has been established. In case a dominant pattern [of usage] has been fairly evidenced, the party relying on the usage is entitled ... to go to the trier of fact on the question of whether such dominant pattern has been incorporated into the agreement.

Id. § 490:1-205, Comment 9. Summers and White write that a usage, under the language of 1-205(2), need not be “certain and precise” to fit within the definition of “any practice or method of dealing.” White & *794Summer, supra, § 3-3 at 87. The manner in which the usage of price protection was carried out was presented with sufficient precision to allow the jury to calculate damages at $220,800.

VI

Waiver Or Course Of Performance

Course of performance under the Code is the action of the parties in carrying out the contract at issue, whereas course of dealing consists of relations between the parties prior to signing that contract. Evidence of the latter was excluded by the District Judge; evidence of the former consisted of Shell’s price protection of Nanakuli in 1970 and 1971. Shell protested that the jury could not have found that those two instances of price protection amounted to a course of performance of its 1969 contract, relying on two Code comments. First, one instance does not constitute a course of performance. “A single occasion of conduct does not fall within the language of this section. . . . ” Haw.Rev. Stat. § 490:2-208, Comment 4. Although the Comment rules out one instance, it does not further delineate how many acts are needed to form a course of performance. The prior occasions here were only two, but they constituted the only occasions before 1974 that would call for such conduct. In addition, the language used by a top asphalt official of Shell in connection with the first price protection of Nanakuli indicated that Shell felt that Nanakuli was entitled to some form of price protection. On that occasion in 1970 Blee, who had negotiated the contract with Nanakuli and was familiar with exactly what terms Shell was bound to by that agreement, wrote of the need to “bargain” with Nanakuli over the extent of price protection to be given, indicating that some price protection was a legal right of Nanakuli’s under the 1969 agreement.

Shell’s second defense is that the Comment expresses a preference for an interpretation of waiver.

3. Where it is difficult to determine whether a particular act merely sheds light on the meaning of the agreement or represents a waiver of a term of the agreement, the preference is in favor of “waiver” whenever such construction, plus the application of the provisions on the reinstatement of rights waived . . ., is needed to preserve the flexible character of commercial contracts and to prevent surprise or other hardship.

Id., Comment 3. The preference for waiver only applies, however, where acts are ambiguous. It was within the province of the jury to determine whether those acts were ambiguous, and if not, whether they constituted waivers or a course of performance of the contract. The jury’s interpretation of those acts as a course of performance was bolstered by evidence offered by Shell that it again price protected Nanakuli on the only two occasions of post-1974 price increases, in 1977 and 1978.31

VII

Express Terms As Reasonably Consistent With Usage In Course of Performance

Perhaps one of the most fundamental departures of the Code from prior contract law is found in the parol evidence rule and the definition of an agreement between two parties. Under the U.C.C., an agreement goes beyond the written words on a piece of paper. “ ‘Agreement’ means the bargain of the parties in fact as found in their language or by implication from other circumstances including course of dealing or usage of trade or course of performance as provided in this chapter (sections 490:1-205 and 490:2-208).” Id. *795§ 490:1-201(3). Express terms, then, do not constitute the entire agreement, which must be sought also in evidence of usages, dealings, and performance of the contract itself. The purpose of evidence of usages, which are defined in the previous section, is to help to understand the entire agreement.

[Usages are] a factor in reaching the commercial meaning of the agreement which the parties have made. The language used is to be interpreted as meaning what it may fairly be expected to mean to parties involved in the particular commercial transaction in a given locality or in a given vocation or trade.... Part of the agreement of the parties ... is to be sought for in the usages of trade which furnish the background and give particular meaning to the language used, and are the framework of common understanding controlling any general rules of law which hold only when there is no such understanding.

Id. § 490:1-205, Comment 4. Course of dealings is more important than usages of the trade, being specific usages between the two parties to the contract. “[C]ourse of dealing controls usage of trade.” Id. § 490:1 — 205(4). It “is a sequence of previous conduct between the parties to a particular transaction which is fairly to be regarded as establishing a common basis of understanding for interpreting their expressions and other conduct.” Id. § 490:1-205(1). Much of the evidence of prior dealings between Shell and Nanakuli in negotiating the 1963 contract and in carrying out similar earlier contracts was excluded by the court.32

A commercial agreement, then, is broader than the written paper and its meaning is to be determined not just by the language used by them in the written contract but “by their action, read and interpreted in the light of commercial practices and other surrounding circumstances. The measure and background for interpretation are set by the commercial context, which may explain and supplement even the language of a formal or final writing.” Id., Comment 1. Performance, usages, and pri- or dealings are important enough to be admitted always, even for a final and complete agreement; only if they cannot be reasonably reconciled with the express terms of the contract are they not binding on the parties. “The express terms of an agreement and an applicable course of dealing or usage of trade shall be construed wherever reasonable as consistent with each other; but when such construction is unreasonable express terms control both course of dealing and usage of trade and course of dealing controls usage of trade.” Id. § 490:1-205(4).

Of these three, then, the most important evidence of the agreement of the parties is their actual performance of the contract. Id. The operative definition of course of performance is as follows: “Where the contract for sale involves repeated occasions for performance by either party with knowledge of the nature of the performance and opportunity for objection to it by the other, any course of performance accepted or acquiesced in without objection shall be relevant to determine the meaning of the agreement.” Id. § 490:2-208(1). “Course of dealing ... is restricted, literally, to a sequence of conduct between the parties previous to the agreement. However, the provisions of the Act on course of performance make it clear that a sequence of conduct after or under the agreement may have equivalent meaning (Section 2-208).” Id. 490:1-205, Comment 2. The importance of evidence of course of performance is explained: “The parties themselves know best what they have meant by their words of agreement and their action under that agreement is the best indication of what that meaning was. This section thus rounds out the set of factors which determines the meaning of the ‘agreement’ ...” Id. § 490:2-208, Comment 1. “Under this section a course of performance is always relevant to deter*796mine the meaning of the agreement.” Id., Comment 2.33

Our study of the Code provisions and Comments, then, form the first basis of our holding that a trade usage to price protect pavers at times of price increases for work committed on nonescalating contracts could reasonably be construed as consistent with an express term of seller’s posted price at delivery. Since the agreement of the parties is broader than the express terms and includes usages, which may even add terms to the agreement,34 and since the commercial background provided by those usages is vital to an understanding of the agreement, we follow the Code’s mandate to proceed on the assumption that the parties have included those usages unless they cannot reasonably be construed as consistent with the express terms.

Federal courts usually have been lenient in not ruling out consistent additional terms or trade usage for apparent inconsistency with express terms. The leading case on the subject is Columbia Nitrogen Corp. v. Royster Co., 451 F.2d 3 (4th Cir. 1971). Columbia, the buyer, had in the past primarily produced and sold nitrogen to Royster. When Royster opened a new plant that produced more phosphate than it needed, the parties reversed roles and signed a sales contract for Royster to sell excess phosphate to Columbia. The contract terms set out the price that would be charged by Royster and the amount to be sold. It provided for the price to go up if certain events occurred but did not provide for price declines. When the price of nitrogen fell precipitously, Columbia refused to accept the full amount of nitrogen specified in the contract after Royster refused to renegotiate the contract price. The District Judge’s exclusion of usage of the trade and course of dealing to explain the express quantity term in the contract was reversed. Columbia had offered to prove that the quantity set out in the contract was a mere projection to be adjusted according to market forces. Ambiguity was not necessary for the admission of evidence of usage and prior dealings.35 Even though the lengthy contract was the result of long and careful negotiations and apparently covered every contingency, the appellate court ruled that “the test of admissibility is not whether the contract appears on its face to be complete in every detail, but whether the proffered evidence of course of dealing and trade usage reasonably can be construed as consistent with the express terms of the agreement.” Id. at 9. The express quantity term could be reasonably construed as consistent with a usage that such terms would be mere projections for several reasons: 36 *797(1) the contract did not expressly state that usage and dealings evidence would be excluded; (2) the contract was silent on the adjustment of price or quantities in a declining market; (3) the minimum tonnage was expressed in the contract as Products Supplied, not Products Purchased; (4) the default clause of the contract did not state a penalty for failure to take delivery; and (5) apparently most important in the court’s view, the parties had deviated from similar express terms in earlier contracts in times of declining market. Id. at 9-10. As here, the contract’s merger clause said that there were no oral agreements. The court explained that its ruling “reflects the reality of the marketplace and avoids the overly legalistic interpretations which the Code seeks to abolish.” Id. at 10. The Code assigns dealing and usage evidence “unique and important roles” and therefore “overly simplistic and overly legalistic interpretation of a contract should be shunned.” Id. at 11.

Usage and an oral understanding led to much the same interpretation of a quantity term specifying delivery of 500 tons of stainless-steel solids in Michael Schiavone & Sons, Inc. v. Securalloy Co., 312 F.Supp. 801 (Conn. 1970). In denying summary judgment for plaintiff-buyer, the court ruled that defendant-seller could attempt to prove that the quantity term was modified by an oral understanding, in line with a trade usage, that seller would only supply as many tons as he could, with 500 tons the upper limit. The court reasoned that an additional term with a lesser effect than total contradiction or negation of a contract term can be a consistent term and “[evidence that the quantity to be supplied by defendant was orally understood to be up to 500 tons cannot be said to be inconsistent with the terms of the written contract which specified the quantity as ‘500 Gross Ton.’ ” Id. at 804.37

The Tenth Circuit in Amerine National Corp. v. Denver Feed Co., 493 F.2d 1275 (10th Cir. 1974), found that the warranty that Amerine, the seller, would provide turkeys was not breached by delivery of “H&N” turkeys instead of “Amerine”. In light of Amerine’s prior dealings and a trade usage that “Amerine” or any trade-name turkey simply meant any turkeys sold by that manufacturer, not a particular kind or breed of turkey, any agreement to provide turkeys did not oblige Amerine to provide only its own strain of turkeys.

*798The Fifth Circuit, in a carefully reasoned opinion, Chase Manhattan Bank v. First Marion Bank, 437 F.2d 1040 (5th Cir. 1971), reversed the lower court’s refusal to allow usage and dealing evidence that seemingly contradicted the express term. The standby agreement in Chase stated that none of the banks would unilaterally demand that the debtor pay up or sell the collateral stock unless all creditors agreed to the sale. The subordination agreement — signed after the standby agreement, incorporated into it, and by its terms to last a maximum of 18 months — provided that the other creditor banks were subordinated to the extent of $1 a share to Chase because of its further loan to the debtor. The usage of the trade and the course of dealing between the parties, ruled inadmissible by the lower court, were that the parties had not intended to limit the duration of the subordination agreement to 18 months.38 The appellate court reversed, directing that evidence of usage and dealings be admitted because it “merely delineates a commercial backdrop for intelligent interpretation of the agreement,” without “delimitpng] a particular party’s intent, except insofar as it reveals that some ascribed intent might be ludicrous in the commercial world.” Id. at 1046.39 The court wrote, “In providing for the admission of such evidence, the Code manifests the law’s recognition of the fact that perception is conditioned by environment: unless a judge considers a contract in the proper commercial setting, his view is apt to be distorted or myopic, increasing the probability of error.” Id. The court added, “If, in light of banking practices in problem loan situations and Chase’s dealings with First Marion, an unsecured loan to the [debtor] would have seemed unreasonable, ambiguities arise within the subordination provision.” Id. at 1047. Usage and dealings evidence here would “permit analysis of the written agreement in the proper commercial setting. Such evidence might disclose ambiguities within the provisions of the agreement. . . . ” Id. For a court to use usage evidence to better understand express terms does not mean it is allowing the written instrument to be contradicted; the use of such evidence “simply places the court in the position of the parties when they made the contract, and enables it to appreciate the force of the words they used in reducing it to writing.” Id. at 1048. “The object of rules of construction generally, and of parol evidence particularly, is to ascertain the intention of the parties.” Id40 It is noteworthy that in Chase, the contradiction was total, not the partial exception Nanakuli argues here.

The Fourth Circuit has been similarly liberal in admitting parol evidence to contradict express terms. After its pacesetting decision in Columbia Nitrogen, supra, it held in Brunswick Box Co. v. Coutinho, Caro & Co., 617 F.2d 355, 360-61 (4th Cir. 1980), that “the Parol Evidence Rule is not a bar to the introduction of extrinsic evi*799dence as to the intention of the parties in the use of the term ‘F.A.S. Norfolk, Virginia’, in the written agreement,” even though “the term, on its face, is unambiguous.” The appeals court, therefore, reversed the lower court’s exclusion ruling and remanded for trial. The term “F.A.S.” is defined by the Code as meaning delivered to the buyer alongside the vessel by the seller, which was also the usage in the port. The plaintiff-seller, however, argued that the dealings between the parties leading to the contract, the actual agreement of the parties, and their course of performance of the contract were that the seller would unload on the dock area. The Code does not bar such evidence “simply because a contract appears on its face to be complete, ...” and therefore plaintiff-seller should have been allowed to show that the parties agreed that the seller would unload the material on the dock area rather than alongside the vessel. Id. at 359.

The Ninth Circuit’s most recent reference to the U.C.C.’s parol evidence rule was in a similar case, Board of Trade of San Francisco v. Swiss Credit Bank, 597 F.2d 146 (9th Cir. 1979), in which this court considered the meaning of an express term in a letter of credit requiring presentment of a “full set clean on board bills of lading.” When the components for the electronic calculators were sent by air, the bank protested that ocean shipment was required by the trade usage as to the express term. Although the U.C.C.’s definition of bills of lading includes airbills, this court upheld the bank’s right to prove a trade usage that in essence contradicted the Code’s broad definition of bills of lading by showing that only ocean bills of lading were allowable. The summary judgment for plaintiff was reversed and the case remanded for trial on whether the bank’s dishonor of the documentary letter of credit was wrongful. The court cited the California Supreme Court, “The test of admissibility of extrinsic evidence to explain the meaning of a written instrument is not whether it appears to the court to be plain and unambiguous on its face, but whether the proffered evidence is relevant to prove a meaning to which the language of the instrument is reasonably susceptible.” Id. at 148-49.41

Numerous state courts have interpreted their own state’s versions of the Code in line with the weight of federal authority on the U.C.C. to admit freely evidence of additional terms, usages, and prior dealings and harmonize them in most instances with apparently contradictory express terms. The only Hawaiian case on the subject dealt with the parol evidence rule in dicta. Cosmopolitan Financial Corp. v. Runnels, 625 P.2d 390 (Haw.App.1981), cert. denied, (Hawaii S.Ct. 4/20/81). The rule did not bar evidence of an additional oral term between an officer of a financial institution and the guarantors of a promissory note that the latter would not be liable in the event the promissor defaulted. Because the additional term was fraudulent and thus invalidated the entire agreement, “the parol evidence rule was inapplicable.” Id. at 396. The court’s discussion of the rule, however, is pertinent as an indication of the approach it would take in this case. It wrote, id. at 395:

Historically, in an action to determine the parties’ contractual rights under an agreement, the court’s only inquiry would *800center around whether the written agreement was a total integration of the parties’ intent. If so, absent evidence of mistake or fraud, the rule barred introduction of any extrinsic evidence that varied or altered the terms. . . . However, since the advent of the adoption of the [Code] in practically every state, rigid adherence to the exclusionary effects of the parol evidence rule has seen a relaxation of its application by the courts in many jurisdictions. This has been largely attributed to a combination of the U.C. C.’s intent to facilitate the flow in business and commercial transactions, and the widespread use of standard business forms to evidence the existence of contractual relationships between parties. For example, article 2 of the U.C.C. permits the court to consider a far wider range of extrinsic evidence to discern the intent of the parties than has been permitted under contract law. . . . [W]e think that expansion of the liberal approach toward the receipt of extrinsic evidence, in the face of the proliferation of standard form contracts and commercial paper, gives the courts a wider insight into the real intent of the parties.

The district judge, “in his refusal to bar evidence of the circumstances surrounding the transaction, was applying this modern principle,” which was the “same view adopted” in a law review article cited by the court:

As between immediate parties, however, all evidence whether written or oral, whether of conditions precedent or subsequent, should be admitted to determine what the parties understood the true contractual relationship to be. Any inherent improbability, such as a contradiction between what allegedly was agreed upon and what was signed will naturally affect the weight to be accorded such evidence, but procedural wrangles can be avoided by allowing the fact finder to hear all the evidence which either party wishes to bring to bear.

Id. at 396 (citing E. R. Jordan, “Just Sign Here—It’s Only a Formality”: Parol Evidence in the Law of Commercial Paper, 13 Ga.L.Rev. 53, 95 (1978)).

The first major state case on the Code’s treatment of trade usage was Provident Tradesmens Bank & Trust Co. v. Pemberton, 196 Pa.Super. 180, 173 A.2d 780, 783-84 (1961). Plaintiff bank, which had financed the purchase of a new car on which the owner defaulted, obtained judgment against the buyer, as well as against the president of the automobile dealer as surety. The surety agreement signed by the dealer’s president contained a written waiver of “all notices whatsoever in respect to this agreement” and further provided that the surety’s liability was “absolute and unconditional and shall not be affected or released by reason of any action taken by the Bank which is hereby consented or agreed to.” The prior course of dealings between the parties and the trade usage, however, were that the bank or other financing institution would notify the dealer when the car was damaged in a collision and the insurance on the car was cancelled. Despite the apparent contradiction between the practice of notification and the express term of no notice, the court held that the bank was obliged to give notice, even though the prior dealings that notice was provided was pursuant to an express term in previous contracts calling for notification in the event of a lapse or cancellation of insurance. However, the failure of the bank’s printed form surety agreement to refer specifically to the custom of notifying a dealer when insurance lapsed meant that the trade usage was not carefully negated as the Code demands.

In A & G Construction Co. v. Reid Brothers Logging Co., 547 P.2d 1207 (Alaska 1976), the dispute between a highway paving contractor and a hot rock supplier was over whether the contractor was liable for all the hot rock delivered to its stockpile or only for that weighed by the state just before it was actually put on the road, and after it was mixed with oil at A & G’s plant. Quantities of hot rock had been lost between the stockpile and the road. The contract term seemed unambiguous: “Payment for the materials will be on the basis *801of State accepted scale ticketed tonnage,” which weighing took place after it was delivered to the stockpile. However, the contract elsewhere referred to the state’s “stockpile payment price,” and the trade usage evidence was that the regular method of delivery was to the contractor’s stockpile. Id. at 1212 n. 3. The Court thus held the supplier should be paid for all hot rock delivered to the stockpile, despite the express reference to the state-scales weighing.

Despite contract specifications that air-conditioning cooling “[c]apacities shall not be less than indicated,” a trade usage that “reasonable variations in cooling capacities were considered to comply with specifications” was admissible in a dispute over a 6% deficiency in cooling capacities. Modine Manufacturing Co. v. North East Independent School District, 503 S.W.2d 833, 838 (Tex.Civ.App.1974). A jury verdict against the manufacturer after usage was excluded was reversed. Id. at 837—40. Part of the basis of the holding was the fact that the contract did not carefully negate the applicability of any trade usage. Id. at 839. Accord, Kenneth Reed Construction Corp. v. United States, 475 F.2d 583 (Ct.Cl.1973) (Provision that “forms shall be true to line and grade” permitted reasonable deviations in line and grade of forms). Because an express term in a written contract that clothing would be delivered in “June— Aug.” could be construed as reasonably consistent with a usage of the trade that all the merchandise could not be shipped in August, summary judgment for the manufacturer-seller was reversed. In Warren’s Kiddie Shoppe, Inc. v. Casual Slacks, Inc., 120 Ga.App. 578, 171 S.E.2d 643 (1969), seller relied on the express term as meaning that he was not in breach for making the first shipments around August 15, and would not be in default if he shipped all of the merchandise by August 31. However, the usage was that “the phrase ‘June-Aug.’ has a definite meaning in the teenage clothing trade — i. e., the largest shipments in June with a substantially similar shipment in July, and the balance . . . approximately twenty percent, ... to arrive in August.” Id. at 644. As in the instant case, one might wonder why the parties did not include with more precision the manner in which the contract was to be carried out; however, the inclusion or exclusion of such terms is often also part of the usage of the trade. In an analogous situation, an additional term, although in apparent conflict with the contract’s express delivery date, could modify the written term, and thus judgment for the seller was reversed in MacGregor v. McReki, Inc., 30 Colo.App. 196, 494 P.2d 1297 (1971). Although the express term merely read that “Buyer understood that seller would try to ship by approximately the first of December,” the appellate court ordered the lower court to admit evidence of an oral agreement by seller to a “delivery date of no later than the end of the second week of December.” Id. at 1298-99. That agreement did not contradict or negate the express term: “To be inconsistent, the offered evidence must contradict or negate the written terms.” Id. at 1299.42

Probably the two leading cases that have rejected usage evidence as inconsistent with express terms are Southern Concrete Services, Inc. v. Mableton Contractors, Inc., 407 *802F.Supp. 581 (N.D.Ga.1975), aff’d, 569 F.2d 1154 (5th Cir. 1978) (unpublished opinion), and Division of Triple T Service, Inc. v. Mobil Oil Corp., 60 Misc.2d 720, 304 N.Y. S.2d 191 (Sup.Ct.1969). In Southern Concrete the District Court, distinguishing its facts from those in Columbia Nitrogen, supra, held that evidence of a trade usage and an agreement to additional terms was not admissible. The usage allegedly was that contract quantity specifications were not mandatory on either buyer or seller. The court acknowledged that U.C.C. § 2-202 “was meant to liberalize the common law parol evidence rule to allow evidence of agreements outside the contract, without a prerequisite finding that the contract was ambiguous” and “requires that contracts be interpreted in light of the commercial context in which they were written and not by the rules on legal construction.” Southern Concrete, supra, at 582-83. Nevertheless, the court held, the express quantity term in the contract and the usage could not be construed as reasonably consistent. “A construction which negates the express terms of the contract by allowing unilateral abandonment of its specifications is patently unreasonable.” Id. at 585. The court’s attempt to differentiate its facts from those in Columbia Nitrogen was unsuccessful; the distinctions discussed were very minor. The difference between the two results should depend less on such subtle variations in contract language and more on the strength of the usage evidence and whether the parties are or should be aware of the usage and thus should be bound by it. The court in Southern Concrete acknowledged that Columbia Nitrogen is not the only case at odds with its holding that a usage that quantities are projections cannot modify a seemingly unambiguous quantity term. Southern Concrete, supra, at 585-86.

The other leading case cited by Shell is a New York case, Triple T, supra. Because the express term of the franchise agreement gave either party the right to terminate 90-days’ notice, the court refused to find as reasonably consistent with that term a usage of the trade that a gasoline franchisor could only terminate a dealer for “cause”. “[T]he express terms of the contract cover the entire area of termination and negate plaintiff’s argument that the custom or usage in the trade implicitly adds the words ‘with cause’ in the termination clause. The contract is unambiguous and no sufficient basis appears for a construction which would insert words to limit the effect of the termination clause.” Id. at 203. The court then held that only consistent usagtes are admissible, which is an incorrect reading of the Code. Usage is always admissible, even though the express term controls in the event of inconsistency, which is a jury question.

Higher New York courts have not been as quick to reject evidence of additional terms for inconsistency as was the Supreme Court in Triple T in rejecting usage evidence. In Hunt Foods & Industries, Inc. v. Doliner, 26 App.Div.2d 41, 270 N.Y.S.2d 937 (1966), the inconsistency with the express term was no greater. When the buyer sued for specific performance of an unconditional option to purchase the stock of the seller corporation, the seller persuaded the court that, despite the unconditional nature of the option, there was an additional understanding that the option would be used only if the seller solicited an outside offer. Negotiations for the acquisition of the corporation’s assets recessed for several weeks after the price was agreed upon but other terms were not; the option was given, the seller testified, because of the buyer’s concern that he would use the offer to solicit a higher bid from a third party. The Appellate Division said, “To be inconsistent, the term must contradict or negate a term of the writing. A term or condition which has a lesser effect is provable.” Id. at 940. “It is not sufficient that the existence of the condition is implausible. It must be impossible.” Id. The jury then determines whether or not it is reasonably consistent with express terms. Similarly, in Ciunci v. Wella Corp., 26 App.Div.2d 109, 271 N.Y. S.2d 317 (1966), the Appellate Division found an additional term consistent with the express terms and reversed summary judgment for the manufacturer on the issue *803of liability for injury to a model’s ear from a hair treatment. The model’s testimony was that, when she was given a free hair treatment, she was told she had to sign a card releasing the student beautician and the manufacturer from “any and all liability for any damages and/or injuries” caused by the treatment or the products because the manufacturer did not want to be responsible for any damage to her hair. This narrower limitation of liability could be seen by a jury as reasonably consistent with the apparently blanket express waiver of all liability in the contract, thus allowing her to attempt to recover for injuries to her ear.

Directly in conflict with the holding in Triple T is that of Warrick Beverage Corp. v. Miller Brewing Co., 170 Ind.App. 114, 352 N.E.2d 496 (1976). The contract to establish a franchise for a beer dealership had no term binding the brewer as to the number of such dealerships, yet usage was admissible to establish that “it was customary in the beer industry to establish a single rather than dual distributorships.” Id. at 501. That evidence of a single dealership limitation for each area was fully consistent with and “would in no way contradict the [express] terms, ...” the court held. Id. That holding, in contrast to the holding in Triple T, is in keeping with the policy of allowing the introduction of such a usage if it is not a total negation of the contract term but simply a partial exception to it.43

Some guidelines can be offered as to how usage evidence can be allowed to modify a contract.44 First, the court must allow a check on usage evidence by demanding that it be sufficiently definite and widespread to prevent unilateral post-hoc revision of contract terms by one party. The Code’s intent is to put usage evidence *804on an objective basis. J. H. Levie, Trade Usage and Custom Under the Common Law and the Uniform Commercial Code, 40 N.Y. U.L.Rev. 1101 (1965), states:

When trade usage adds new terms to cover matters on which the agreement is silent the court is really making a contract for the parties, even though it says it only consulted trade usage to find the parties’ probable intent. There is nothing wrong or even unusual about this practice, which really is no different from reading constructive conditions into a contract. Nevertheless the court does create new obligations, and perhaps that is why the courts often say that usage . .. must be proved by clear and convincing evidence.

Id. at 1102. Although the Code abandoned the traditional common law test of nonconsensual custom and views usage as a way of determining the parties’ probable intent, id. at 1106-07, thus abolishing the requirement that common law custom be universally practiced, trade usages still must be well settled, id. at 1113. Columbia Nitrogen, supra, has been criticized as allowing the introduction of evidence of a usage that was not proved to be sufficiently well-established and that was not clearly applicable to the particular facts of the detailed contract between Royster and Columbia Nitrogen. Kirst, Usage of Trade and Course of Dealing: Subversion of the UCC Theory, 1977 Law Forum, 811; Note, Commercial Law — Course of Dealing and Usage of Trade Affect Express Terms, 1973 Wisc.L. Rev. 934, 940-43.

Evidence of a trade usage does not need to be protected against perjury because, as one commentator has written, “an outside standard does exist to help judge the truth of the assertion that the parties intended the usage to control the particular dispute: the existence and scope of the usage can be determined from other members of the trade.” Kirst, supra, at 839. Kirst sets out guards on jury determination of usage evidence:

Questions of the parties’ intentions concerning an asserted trade usage or course of dealing will not always require a jury determination. If the evidence fails to show a practice is regularly observed, the judge can exclude the evidence because it does not show a course of dealing or usage of trade as defined in the Code. If the members of the trade confirm an actual usage but do not support the assertion that the usage applies to the particular facts in litigation, the judge will exclude evidence of the usage as irrelevant. If the parties used new and different language to convey their agreed intention to abandon the past practice, the court will recognize that practice under the old language is irrelevant to the contract containing the new language and, consequently, will exclude the evidence.

In Columbia Nitrogen, supra, the court should have examined the relationship of the usage to the facts of the case, for example, by determining whether any of the contracts that had been treated as only “fair estimates” in the past were the detailed result of long negotiations, as was the contract in Columbia Nitrogen. That contract

was the result of extensive negotiations for the sale of part of Royster’s output from a major new facility covering sales for three years. If contracts of similar detail, covering similar time periods, and negotiated with similar care were regarded as estimates by others in the trade, then the court should have admitted the asserted usage of trade. If the contracts treated as estimates, however, were always form contracts, or short term contracts, or otherwise substantially different, then the court should have excluded the evidence as irrelevant to the dispute.

Id. at 845. That formulation of relevance of the usage evidence seems a fair one to follow in this case. Here the evidence was overwhelming that all suppliers to the asphaltic paving trade price protected customers under the same types of circumstances. Chevron’s contract with H.B. was a similar long-term supply contract between a buyer *805and seller with very close relations, on a form supplied by the seller, covering sales of asphalt, and setting the price at seller’s posted price, with no mention of price protection. The same commentator offers a second guideline:

Because the stock printed forms cannot always reflect the changing methods of business, members of the trade may do business with a standard clause in the forms that they ignore in practice. If the trade consistently ignores obsolete clauses at variance with actual trade practices, a litigant can maintain that it is reasonable that the courts also ignore the clauses. Similarly, members of a trade may handle a particular subset of commercial transactions in a manner consistent with written terms because the writing cannot provide for all variations or contingencies. Thus, if the trade regards an express term and a trade usage as consistent because the usage is not a complete contradiction but only an occasional but definite exception to a written term, the courts should interpret the contract according to the usage.

Kirst, supra, at 824. Levie, supra, at 1112, writes, “Astonishing as it will seem to most practicing attorneys, under the Code it will be possible in some cases to use custom to contradict the written agreement. . . . Therefore usage may be used to ‘qualify’ the agreement, which presumably means to ‘cut down’ express terms although not to negate them entirely.” Here, the express price term was “Shell’s Posted Price at time of delivery.” A total negation of that term would be that the buyer was to set the price. It is a less than complete negation of the term that an unstated exception exists at times of price increases, at which times the old price is to be charged, for a certain period or for a specified tonnage, on work already committed at the lower price on nonescalating contracts. Such a usage forms a broad and important exception to the express term, but does not swallow it entirely. Therefore, we hold that, under these particular facts, a reasonable jury could have found that price protection was incorporated into the 1969 agreement between Nanakuli and Shell and that price protection was reasonably consistent with the express term of seller’s posted price at delivery.

VIII

Good Faith In Setting Price

Nanakuli offers an alternative theory why Shell should have offered price protection at the time of the price increases of 1974. Even if price protection was not a term of the agreement, Shell could not have exercised good faith in carrying out its 1969 contract with Nanakuli when it raised its price by $32 effective January 1 in a letter written December 31st and only received on January 4, given the universal practice of advance notice of such an increase in the asphaltic paving trade. The Code provides, “A price to be fixed by the seller or by the buyer means a price for him to fix in good faith,” Haw.Rev.Stat. § 490:2-305(2). For a merchant good faith means “the observance of reasonable commercial standards of fair dealing in the trade.” Id. 490:2-103(l)(b). The comment to Section 2-305 explains, “[I]n the normal case a ‘posted price’ . . . satisfies the good faith requirement.” Id., Comment 3. However, the words “in the normal case” mean that, although a posted price will usually be satisfactory, it will not be so under all circumstances. In addition, the dispute here was not over the amount of the increase — that is, the price that the seller fixed — but over the manner in which that increase was put into effect. It is true that Shell, in order to observe the good faith standards of the trade in 1974, was not bound by the practices of aggregate companies, which did not labor under the same disabilities as did asphalt suppliers in 1974. However, Nanakuli presented evidence that Chevron, in raising its price to $76, gave at least six weeks’ advance notice, in accord with the long-time usage of the asphaltic paving trade. Shell, on the other hand, gave absolutely no notice, from which the jury could have concluded that Shell’s manner of carrying out the price increase of 1974 did not conform to commercially reasonable standards. In *806both the timing of the announcement and its refusal to protect work already bid at the old price, Shell could be found to have breached the obligation of good faith imposed by the Code on all merchants. “Every contract or duty within this chapter imposes an obligation of good faith in its performance or enforcement,” id. § 490:1-203, which for merchants entails the observance of commercially reasonable standards of fair dealing in the trade. The Comment to 1-203 reads:

This section sets forth a basic principle running throughout this Act. The principle involved is that in commercial transactions good faith is required in the performance and enforcement of all agreements or duties. Particular applications of this general principle appear in specific provisions of the Act. ... It is further implemented by Section 1-205 on course of dealing and usage of trade.

Id. § 490:1-203, Comment. Chevron’s conduct in 1974 offered enough relevant evidence of commercially reasonable standards of fair dealing in the asphalt trade in Hawaii in 1974 for the jury to find that Shell’s failure to give sufficient advance notice and price protect Nanakuli after the imposition of the new price did not conform to good faith dealings in Hawaii at that time.

Because the jury could have found for Nanakuli on its price protection claim under either theory, we reverse the judgment of the District Court and reinstate the jury verdict for Nanakuli in the amount of $220,800, plus interest according to law.

REVERSED AND REMANDED WITH DIRECTIONS TO ENTER FINAL JUDGMENT.

KENNEDY, Circuit Judge,

concurring specially:

The case involves specific pricing practices, not an allegation of unfair dealing generally. Our opinion should not be interpreted to permit juries to import price protection or a similarly specific contract term from a concept of good faith that is not based on well-established custom and usage or other objective standards of which the parties had clear notice. Here, evidence of custom and usage regarding price protection in the asphaltic paving trade was not contradicted in major respects, and the jury could find that the parties knew or should have known of the practice at the time of making the contract. In my view, these are necessary predicates for either theory of the case, namely, interpretation of the contract based on the course of its performance or a finding that good faith required the seller to hold the price. With these observations, I concur.