In this patent infringement case, Linear Technology Corporation (“LTC”) sued Mi-crel, Inc. (“Micrel”) for infringement of U.S. Patent No. 4,755,741 and Reexamination Certificate B1 4,755,741 (collectively, “the '741 patent”), directed towards adaptive transistor drive circuitry. Following a bench trial limited solely to Mi-crel’s invalidity defense, the district court held the '741 patent invalid due to the on-sale bar of 35 U.S.C. § 102(b).
Linear Tech. Corp. v. Micrel, Inc.,
On appeal, LTC challenges the district court’s judgment of invalidity under section 102(b), and Micrel cross-appeals several of the district court’s evidentiary rulings excluding letters proffered by Micrel that purportedly support its argument that an invalidating offer for sale occurred.
In this appeal, we must decide whether the facts found by the district court support its conclusion that an invalidating sale or offer for sale occurred before the statutory critical date, which in this case is November 18, 1985. In
Group One, Ltd. v. Hallmark Cards, Inc.,
I
BACKGROUND
LTC, the assignee of the '741 patent, designs, manufactures, and sells linear integrated circuits for use in a variety of applications, including telecommunications, cellular phones, and computers. Carl Nelson, an LTC employee and the named inventor of the '741 patent, conceived the invention while developing a new silicon chip for LTC, a chip that eventually became known as the LT1070. The patented' invention involves switching regulator cir *1044 cuitry that provides regulated voltages or currents to electrical devices. According to Nelson, the LT1070 chip is a functioning version of the invention claimed in the '741 patent.
In the months and weeks leading up to the official release date of the LT1070 chip on November 18, 1985, LTC engaged in extensive pre-release marketing activity designed to generate commercial interest in the chip both in the United States and abroad. During this time, LTC employed independent sales representative firms to sell its products in the domestic market. LTC also sold its products in Europe through its independent European distributors, who served a dual role: as customers they purchased products from LTC for resale, while as sales representatives they marketed and sold those products to end-users in Europe. During the 1980s, LTC used product data sheets as a primary selling tool; these sheets informed customers of LTC product specifications so that the customers could design LTC components into their own products. LTC generated a pre-release preliminary data sheet for the LT1070, which listed many of the salient features of the new chip. Though LTC did not receive the glossy, professionally printed version of the preliminary data sheet from the printer until late October 1985, a yet-more-preliminary form of the data sheet nonetheless made its way into the hands of LTC’s sales representatives before October, and the sales representatives sent the data sheets to potential customers before the critical date.
In July of 1985, LTC conducted a sales conference in Santa Clara, California, to which it invited both its domestic and international sales representatives. Robert Scott, LTC’s Product Marketing Manager in 1985, testified that LTC held the conference “to inform those attending of new and future LTC products, including ... unannounced products which were still being developed but which were not yet available for sale.” Two presentations at the conference discussed the LT1070, and one of the attendees sketched an application circuit diagram of the LT1070 into his notebook from the conference. Scott testified that LTC provided information on the LT1070 so that the attendees could identify and familiarize themselves with customers who might be able to use the LT1070, and he admitted that the sales representatives were allowed to talk to customers about the part, albeit in less detail than for a released product. LTC also pitched the LT1070 in a newsletter that it distributed to all its domestic sales representatives and international distributors on or about November 1, 1985. The newsletter enthused:
[T]he LT1070 promises to be a big hit at accounts facing the decision of whether to ‘make or buy’ their next power supply.... Carl Nelson’s goal in designing the LT1070 series is to provide the user a high performance-easy-to-use-switching regulator requiring very few external components, at a very competitive cost compared to conventional power supply designs. Our first look at samples says the LT1070 series won’t disappoint anyone....
Meanwhile, Hans Zaph, LTC’s Director of International Sales, praised the LT1070 directly to the international distributors, calling it a “brand new switch mode power supply circuit, which will be a true industry first.”
By October 1985, LTC had generated enough excitement that the independent sales representatives were actively promoting the LT1070 to customers, even though the new chip had not yet been released for sale. Indeed, one independent sales representative exhorted his col *1045 leagues in an intra-office memorandum that “WE REALLY MUST START BROADENING OUR ACCOUNT BASE NOW. THE LT1070 CAN BE SOLD ANYWHERE; BEST DAMNED POWER SUPPLY PART IN HISTORY.”
Unsurprisingly in light of its marketing blitzkrieg, LTC received purchase orders from four of its European distributors offering to buy LT1070 chips before the official release date. Following its standard procedure for unreleased products, LTC did not book the four purchase orders for the LT1070. Instead, LTC entered the orders into its computerized order tracking system using what the parties refer to as the “will-advise” procedure. Instead of entering part number, quantity and price terms as they would for an order for a released product, LTC customer service representatives entered the LT1070 orders into a computer using the words WILL ADVISE.” LTC then added the following words to the computer-generated form, “NEW PRODUCT/NOT RELEASED,” WILL ADVISE ON PART # ORDERED-NOT BOOKED” in place of the part number and part description fields in their tracking system. In all but one case, 1 LTC entered “0” in the price field, and “1” in the quantity field regardless of the number of chips actually ordered, because the software required a nonzero number in the quantity field. LTC’s system thus generated an acknowledgement form containing the above-described information, and a customer service representative faxed the acknowledgement to the foreign distributors to notify them of the status of their order. LTC customer service periodically checked the status of the will-advise orders, and after LTC officially released the LT1070, customer service deleted and reentered the will-advise orders as “normal” orders using the newly-minted LT1070 part number, the actual quantity ordered, and the actual price term. This transformation did not require — and LTC never sought — any further communication from the distributors regarding the order, and LTC shipped the LT1070s to the distributors shortly thereafter. From the distributors’ perspective, therefore, the LT1070s arrived without any further action by them other than their pre-release-date submission of a purchase order.
Following their extensive pre-release marketing build-up, LTC officially released the LT1070 for sale on November 18, 1985, exactly one year before it filed the application that eventually issued as the '741 patent. Two reexamination certificates issued on May 14, 1991, and December 26, 1995, as Reexamination Certificates B1 4,755,741 and B2 4,755,741, respectively; both resulted from reexamination requests filed by third parties, including Micrel in the case of the B2 Certificate.
Micrel, a direct competitor of LTC, also makes switching regulator circuitry. On May 9,1994, LTC filed this action claiming that Micrel’s MIC 2172 and MIC 3172 chips infringe the '741 patent. In a motion for summary judgment of invalidity, Micrel argued that LTC offered the LT1070 chip for sale over one year before the November 18, 1986, filing date of the '741 patent, thus triggering the on-sale bar of 35 U.S.C. § 102(b). The district court denied Micrel’s motion, bifurcated the case for a separate early trial on the section 102(b) *1046 issue, and stayed all other issues pending resolution of Mierel’s invalidity defense.
The district court conducted a bench trial on the on-sale bar issue. Following trial, but before the court issued its opinion and judgment, the Supreme Court decided
Pfaff v. Wells Electronics, Inc.,
The district court then turned to “the closer question of whether LTC made a commercial sale or offer for sale in the United States prior to the critical date.”
Id.
at 1123. To decide this issue, whether the LT1070 had been the subject of a commercial offer for sale, the district court applied our
pre-Pfaff
opinion in
RCA Corp. v. Data General Corp.,
The district court carefully documented LTC’s extensive promotional activities pri- or to the critical date. In particular, the court relied on trial testimony indicating “that LTC engaged in substantial preparatory activity prior to the formal release of the LT1070 on November 18, 1985, including seeking feedback from its sales representatives and distributors regarding the pricing of the LT1070 and the publication of data sheets and other promotional materials.”
Linear Tech. Corp.,
Finally, the district court found that LTC “expected its sales force to discuss the LT1070 and its functions with end-user customers,” and that LTC gave the sales representatives information about the LT1070 in order to enable this marketing effort. Id. The court pointed to testimony that “established that data sheets and samples were necessary to offer LTC products for sale to customers,” and that “[sjales personnel at [two of LTC’s domestic sales representative firms] contacted end-user customers and provided them with preliminary LT1070 data sheets in order to seek design-ins for the LT1070.” Id. Discussions between the sales force and customers “resulted in 26 separate requests for LT1070 samples from [LTC’s] domestic and international sales force between July 17, 1985, the date of the sales *1047 conference, and the critical date.... ” Id. Ultimately, the court concluded that “[t]he use of the preliminary data sheets establishes that discussions between LTC sales representatives and end-user customers extended beyond merely nebulous or indefinite discussions about a possible sale,” thus triggering the on-sale bar under RCA Corp. Id.
The court also concluded that LTC’s entry of pre-critical date purchase orders from its international distributors constituted completed sales. The will-advise procedure, under the district court’s theory, was nothing more than a subterfuge obscuring LTC’s actual acceptance of the purchase orders before the critical date. Thus, despite the fact that the orders were not formally booked, the court found that “it is clear that the entry of the orders for the LT1070 into the bookings system constituted a completed sale.” Id. at 1126. Acceptance of these pre-critical date purchase orders under the will-advise procedure thus provided an alternative basis for the court’s ultimate conclusion of invalidity under the on-sale bar.
Based on its extensive findings regarding LTC’s pre-critical date commercialization efforts, as well as findings regarding four pre-critical date purchase orders by LTC’s independent international distributors entered under the will-advise procedure, the district court concluded that LTC had offered for sale and sold the LT1070 prior to the critical date. Id. at 1126-27. Consequently, the district court entered judgment holding the asserted claims of the '741 patent invalid, and LTC appealed, vesting us with jurisdiction pursuant to 28 U.S.C. § 1295(a)(1).
II
A determination that a product was placed on sale prior to the critical date is a conclusion of law based on underlying findings of fact.
Ferag AG v. Quipp, Inc.,
Section 102 of Title 35 of the U.S.Code provides in relevant part that “[a] person shall be entitled to a patent unless ... (b) the invention was ... on sale in this country, more than one year prior to the date of the application for patent in the United States_” To prevail on an on-sale bar defense, an accused infringer must demonstrate by clear and convincing evidence that “there was a definite sale or offer for sale of the claimed invention prior to the critical date, defined as one year prior to the U.S. filing date to which the application was entitled.”
Mas-Hamilton Group v. LaGard, Inc.,
As the district court recognized, the Supreme Court’s opinion in
Pfaff
rejected our previous flexible, totality-of-the-circumstanees on-sale bar analysis in favor of a two-part test. Under
Pfaff’s.
test, the on-sale clock begins to run when two conditions are met: (1) “the product must be the subject of a commercial offer for sale”; and (2) “the invention must be ready for patenting.”
Though the
Pfaff
Court explained at length the “ready for patenting” prong, it provided less guidance on the definition of a “commercial offer for sale.” In the case before it, the inventor had accepted a purchase order memorializing a sale before the critical date, which persuaded the Supreme Court that an offer for sale had been made at some point before that.
Id.
at 67,
In
Group One,
this court applied the offer-for-sale prong of
Pfaff
to a situation in which the patentee engaged in pre-sale negotiations that never resulted in a formal offer. The district court in
Group One
had found that although the negotiations did not rise to the level of an offer for sale under contract law principles, they nevertheless triggered the on-sale bar under our
pre-Pfaff
opinion in
RCA Corp.,
in which we had stated that the on-sale bar may be triggered by “a patentee’s commercial activity which does not rise to the level of a formal ‘offer’ under contract law principles.”
RCA Corp.,
Under
Group One,
interpretation of Pfaff’s commercial offer-for-sale prong thus requires the application of traditional contract law principles to the facts of the particular case. Because the on-sale bar involves questions of patent law requiring national uniformity,
Group One
declined to rely on the law of the particular state in which the transaction occurred, instead holding that the existence of an offer for sale should be analyzed “under the law of contracts as generally understood.”
Group One,
*1049
Though the district court applied
Pfaff,
it did so without the benefit of our opinion in
Group One.
Our task is to consider whether, under the legal principles articulated in
Group One,
the district court’s conclusion that LTC’s activities triggered the on-sale bar may be sustained. As we described above, the district court based its conclusion of invalidity on two alternative grounds: first, that LTC’s active promotion of the LT1070 to customers before the critical date triggered the on-sale bar,
Linear Tech. Corp.,
A
In its conclusions of law, the district court found that “LTC’s sales force was actively promoting the LT1070 to end-user customers” before the critical date, id. at 1125-26, which triggered the on-sale bar. As an initial matter, we agree with the district court that LTC’s activities would have triggered the on-sale bar under the law as enunciated in RCA Corp. LTC’s significant pre-release efforts at commercialization and its targeted advertising to end-users could have triggered the on-sale bar under RCA Corp. even if they did not rise to the level of a contractual offer-for-sale because, as the district court found, they went “beyond merely nebulous or indefinite discussions about a possible sale.” Id. at 1125.
However, the district court’s reliance upon our decision in RCA Corp. for the proposition that promotional activity not rising to the level of a contractual offer-for-sale can trigger the on-sale bar is untenable in light of Group One, which is binding precedent governing this court’s resolution of the case at bar. As we recognized in Group One, Pfaff undercut the RCA Corp. flexible standard and thus removed the pre-release commercialization activity cited by Micrel and the district court as a predicate to a legal conclusion of an on-sale bar. However, our conclusion that the district court applied the incorrect legal standard does not end our inquiry because we may still affirm if the facts reveal an offer for sale under the legal test articulated in Group One.
B
It is of course possible that the facts found by the district court, when analyzed under the test of Group One, support a conclusion of a violation of the on-sale bar. We therefore must assess the four categories of facts upon which the district court focused. Those categories are: (1) LTC’s solicitation of pricing information from its distributors and sales representatives; (2) LTC’s publication of preliminary data sheets and promotional information on the LT1070; (3) LTC’s communications to its sales force in its newsletters and via the sales conference in July of 1985, and the sales representatives’ communications with customers providing them with the LT1070 preliminary data sheets; and (4) sales representatives’ pre-critical date requests for LT1070 samples to give to specific customers. From this evidence, Micrel argues that the district court could infer that invalidating offers for sale had been made. Viewed under general principles of contract law, however, we think that none of these activities constitutes an offer for sale under general principles of contract law.
*1050
The UCC does not define “offer,” so we will look to the common law to guide our inquiry. “An offer is the manifestation of willingness to enter into a bargain, so made as to justify another person in understanding that his assent to that bargain is invited and will conclude it.”
Restatement (Second) of Contracts
§ 24 (1981);
accord
Richard A. Lord,
Williston on Contracts
§ 4:13, at 367 (4th ed.1990) (explaining that “in order for an offer to exist, it must constitute a manifestation communicated to the offeree so as to justify his understanding that by assenting a bargain will be concluded”). Thus, the internal communications between LTC and its sales representatives soliciting information on pricing for the LT1070 cannot constitute an offer for sale to a customer not privy to the communications. But while LTC’s request for pricing information from its United States representatives did not involve communication to a potential buyer, the district court did find that at least one sales representative spoke to potential customers in order to help LTC determine an appropriate price for the chip.
Linear Tech. Corp.,
Similarly, LTC’s mere publication of preliminary data sheets and promotional information for the LT1070 communicates nothing to customers about LTC’s intent, and thus cannot be an offer for sale. Such activities only indicate preparation to place the LT1070 on sale. Preparation alone cannot give rise to an on-sale bar under Group One.
The third category of behavior relied upon by the district court in finding an on-sale bar consists primarily of LTC’s dissemination of advertising materials and other information about the LT1070 to customers. LTC provided this information both orally (through the sales conference) and in writing (through the preliminary data sheets and newsletters). As to the sales conference attended by both domestic sales representatives and the foreign sales representative/distributors, it is clear that LTC designed it to be informational in nature and did not intend to use it to offer the LT1070 for sale. The district court found that LTC “presented the information about the LT1070 so that the sales representatives and international distributors would identify and familiarize themselves with customers who might be able to use the LT1070.”
Linear Tech. Corp.,
*1051
Like the sales conference, the November “Newsline” newsletter extolling the virtues of the LT1070 and LTC’s other products is another form of promotional material,
ie.,
an advertisement. Though primarily directed to domestic sales representatives who never purchased LTC products and to whom LTC did not make offers, the newsletters were also sent to the European distributors. The question therefore becomes whether such promotional materials extended an offer for sale to those distributors. As we noted in
Group One,
“contract law traditionally recognizes that mere advertising and promoting of a product may be nothing more than an invitation for offers, while responding to such an invitation may itself be an offer.”
Group One,
The communications from LTC sales representatives providing customers with preliminary LT1070 data sheets likewise do not reveal the requisite intent to be bound, a
sine qua non
of an offer. Based on testimony from sales representatives, the district court found that LTC sales representatives contacted end-user customers and provided them with data sheets “in order to seek design-ins for the LT1070.”
Linear Tech. Corp.,
The final basis upon which the district court relied to find an offer for sale was the numerous requests for samples made by sales representatives in the days leading up to the November 18 release date. Many of these requests identified specific customers to whom the samples would be provided, suggesting that the sales representatives had communicated with customers about the LT1070. Thus, *1052 the numerous requests show that the LTC sales force was gearing up to offer the LT1070 for sale — but they do not show that the sales force actually made any offers before receiving the samples.
In short, none of the pre-critical date communications between LTC and its sales representatives or between the sales representatives and customers rose to the level of an offer for sale that would form a contract if accepted. Thus, as a matter of law, none of them triggered the on-sale bar under Group One.
C
We turn now to the other basis for the district court’s ultimate conclusion of invalidity under the on-sale bar: its conclusion that LTC sold LT1070s to its European distributors when it entered the pre-critical date purchase orders into its system using the will-advise procedure. General principles of contract law must govern a conclusion that a sale occurred, and under these principles, a completed sale requires an offer, an acceptance, and consideration. We recognize that under the UCC, a valid contract can be found even if the court cannot pinpoint the exact time of its formation, UCC § 2-204(2), but this rule — however laudable in the commercial setting — cannot relieve the court in an on-sale bar setting from ascertaining if a sale, or an offer for sale, has been made before the critical date. The district court evaluated the facts and concluded that such an invalidating sale took place, and we review this conclusion of law de novo. We therefore turn to the facts to determine whether they support the conclusion that at least one invalidating contract for sale was concluded under the will-advise procedure before the critical date.
As an initial matter, it is undisputed that the purchase orders submitted by the foreign distributors were offers to buy the LT1070. The purchase orders contained sufficiently definite terms to create an enforceable contract upon acceptance under the UCC; they included quantity terms and clearly identified the requested product: the LT1070. True, they did not specify a price, but the UCC does not require price terms to create a binding contract. See U.C.C. § 2-305. 3 There can be no doubt that at some point LTC accepted the LT1070 offers, because eventually — and safely after the critical date — it filled the orders. The question is whether LTC accepted them before November 18, 1985, because if so, then it entered into a binding contract to sell the LT1070 that invalidates the '741 patent.
Under general principles of contract law, to accept an offer an offeree must make a manifestation of assent to the offeror.
See Williston on Contracts, supra,
§ 4:1, at 236 (“It is customarily said that mutual assent is essential to the formation of informal contracts, but it must be noted that the mutual assent must be manifested by one party to the other, and except as so manifested, is unimportant.”). Thus, the district court erred as a matter of law in holding that “the entry of the orders for the LT1070 into the bookings system constituted a completed sale,”
Linear Tech. Corp.,
In order to be effective, an acceptance must
objectively
manifest the offeree’s assent.
See Superior Boiler Works, Inc. v. R.J. Sanders, Inc.,
LTC contends that the will-advise ac-knowledgements told the distributors that it “did not accept” the premature purchase orders, a position with which Micrel understandably disagrees. Testimony established that when LTC entered the premature purchase orders under the will-advise procedure, instead of providing the standard booking information for an accepted order, the system-generated confirmation sheet clearly stated “WILL ADVISE,” “NEW PRODUCT, NOT RELEASED,” “WILL ADVISE WILL ADVISE [sic] ON PART# ORDERED-NOT BOOKED.” Customer service at LTC sent the confirmation sheet with these cautionary notes to the foreign distributors. Micrel argues that LTC thereby accepted the purchase orders, thus completing a sale of the LT1070. We think not. In the absence of a course of dealing or other evidence to the contrary, we think that a reasonable offeror in the distributors’ position who received an acknowledgement that — unlike those for the unambiguously-accepted orders — explicitly states that the order for the LT1070 was “NOT BOOKED,” would understand that LTC had not booked and therefore had not accepted the order. 4
When we asked Micrel at oral argument to support its position that the distributors understood the acknowledgement to be an acceptance, Micrel pointed to testimony allegedly showing that the will-advise notation merely notified the distributors that they would be informed of a shipment schedule at a later time. An LTC employee affirmed that the “will-advise” notation in the shipping date column of a post-critical date acknowledgement sent to a European distributor “means that Linear *1054 was going to advise [the distributor] at some point in the future when it would be able to ship the 50 LT1070s that Linear had formally booked on November 19th.” We are not so sanguine as Micrel that this testimony establishes that LTC always meant the “will-advise” notation merely to refer to future advice regarding shipping, particularly as the acknowledgement to which the witness referred was for an order that was “formally booked” and apparently was missing only the shipping date. But even if LTC did attach this idiosyncratic meaning to all will-advise notations, Micrel’s evidence does not answer whether LTC objectively informed the distributors of this fact. Whatever the will-advise notation meant to LTC is not relevant to the section 102(b) inquiry; what it meant to its recipient is critical. Micrel, however, could point to no evidence, either in its brief or at oral argument, that shows what the distributors thought the will-advise notation meant. Indeed, the record is devoid of any evidence, testimonial or otherwise, as to what the distributors actually understood when they received the will-advise confirmations. In the absence of such evidence, we think that the natural reading of ‘WILL ADVISE-NOT BOOKED,” ie. that the orders were not accepted, must prevail.
Micrel urges that we look to the conduct of the parties to show the precritical date existence of a contract, and notes in particular that no further communication from the distributors was required before LTC transformed the “will advise” orders to “normal orders” and shipped them to the buyers. Even if the treatment of the pre-critical date orders were considered as somewhat suspicious, and LTC regarded the “will-advise” notation as nothing more than a surrogate for its subjective intent to accept the orders, nevertheless, it falls on Micrel to prove by clear and convincing evidence that LTC commu- . nicated its subjective intent to the distributors, and this Micrel has failed to do. The fact that the distributors did not submit any further information before receiving the goods does not show that they knew LTC accepted them under the will-advise procedure-it merely demonstrates that they knew LTC accepted the orders at some point. Indeed, under UCC § 2-206(l)(b), shipment itself can be an acceptance, and the distributors could have reasonably assumed that LTC finally accepted their orders when it shipped the LT1070s.
Thus, while we have testimony from LTC as to what it thought the will-advise notation meant internally (ie., just a delay in shipping), and while it is true that the UCC does not require a price term or shipping date to form a valid contract, we are still short a crucial piece of evidence necessary to find a contract: proof of a valid acceptance before the critical date. The missing piece of evidence — the Achilles’ heel in Micrel’s case — is testimony or other evidence as to what the distributors understood when they received their “will advise” acknowledgements. Nor, of equal significance, does the record disclose any evidence suggesting that the distributors were not free to withdraw their offers at any time before their post-critical date acceptance by LTC. Thus, there is no evidence, clear and convincing or otherwise, from which one may draw a conclusion that an acceptance occurred. Without an acceptance, there can be no contract, and hence, under Group One, no on-sale bar. We therefore reverse the district court’s judgment holding the '741 patent invalid under the on-sale bar.
Ill
This brings us to Micrel’s cross-appeal of an evidentiary ruling by the district court excluding certain letters
*1055
discovered in files kept by a sales representative at the now-defunct Rockmore firm, one of the sales representative firms employed by LTC in 1985. On appeal, “[w]e review evidentiary rulings of the district court for abuse of discretion, and interfere with its judgment only if an erroneous ruling prejudiced substantial rights.”
Applied Med. Res. Corp. v. U.S. Surgical Corp.,
Micrel appeals rulings by the district court excluding from evidence three letters addressed from Robert Stenstrom, an LTC sales representative, to customers. The letters discuss the LT1070 and specifically state that an LT1070 preliminary data sheet is enclosed. Micrel sought to admit the letters to show that “LTC, through its sales representatives, sent out data sheets and solicited orders in the United States before the critical date.” The district court excluded the letters on authentication grounds under Fed.R.Evid. 901, and on appeal Micrel argues that the district court erred in so-doing. In light of our decision on the merits of Micrel’s commercialization argument, the letters are not relevant if offered solely for the purpose of showing commercialization that does not rise to the level of an offer for sale. However, assuming arguendo that the letters are relevant to show a contractual offer for sale, after reviewing the district court’s rulings and the letters at issue, we conclude that the court did not abuse its discretion in excluding them based on lack of authentication.
The letters — all of which are dated before the critical date — are not signed, are not on company letterhead, and bear no outward indicia of having ever been mailed to a customer. They were found in files kept by John Moore, a former sales representative of Rockmore. Micrel attempted to authenticate the letters through testimony by Moore and by their putative author, Robert Stenstrom. However, while Stenstrom testified that he sent copies of the LT1070 data sheet to customers, he had no recollection of mailing the letters at issue. Micrel asserts that the copies originally came from a so-called reading file, and notes that only mailed letters were kept in reading files. Ergo, argues Micrel, the letters must have been mailed.
The district court did not find Micrel’s “reading file” evidence persuasive, and neither do we. The problem lies in Micrel’s first assumption: that the letters came from a reading file. First, the Stenstrom letters did not look like letters from a Rockmore reading file because unlike reading file letters, the Stenstrom letters were not three-hole punched. And while Stenstrom, based on his recollection of the business practices of thirteen years before, testified that it was likely that the copies at issue came from a reading file, he admitted that he could not be sure. Nor did Moore’s testimony necessarily provide ade *1056 quate grounds for showing the letters’ original source, because he only had knowledge of reading files in general rather than of these letters specifically. Thus, Moore’s testimony established the general way in which Rockmore stored correspondence in a reading file. However, he did not testify to any personal knowledge regarding the circumstances surrounding the Stenstrom letters, and his only basis for surmising that they came from a reading file was that they were unsigned and not on letterhead. We think — as did the district court — that the fact that the letters were not on letterhead and were unsigned could mean many things, including that the letters were never mailed. Given the general level of uncertainty regarding whether Stenstrom authored — let alone mailed— the letters at issue, the district court was understandably reluctant to admit the letters. The court noted that “it seems to me that there are a lot of problems with the indicia of reliability, and that really is the bottom line through any test with regard to admission.... And to me there’s just too many questions about it.” We agree that given the lingering questions regarding the letters’ authenticity, the district court was well within its discretion when it excluded them because they lacked proper authentication as required by Fed. Rule Evid. 901(a).
IV
For the reasons stated above, we reverse the district court’s judgment of invalidity in favor of Micrel. We leave undisturbed, however, the district court’s evidentiary rulings, thereby affirming the questions raised by Micrel’s cross-appeal. The case is returned to the district court for further proceedings not inconsistent with this opinion.
COSTS
No costs.
AFFIRM-IN-PART, REVERSE-IN-PART, AND REMAND.
Notes
. In one case — a purchase order received from international distributor Neye — LTC entered the actual number ordered (fifty) as well as the price, $5.25. However, this order still contained the will-advise notation in the part number column, and was tracked in the same way as the other will-advise orders.
. As described by the district court, the “process of getting a design engineer at a company to use a component product in one of the engineer's applications is known as a 'design-in.' ”
Linear Tech. Corp.,
. The disputed purchase orders also contained offers to buy other parts that had already been released and were not subject to the will-advise procedure. That LTC could— and did — accept the distributors' offers to buy the released products demonstrates that the purchase orders were "offers to buy” that became binding if accepted.
. We note that the district court did not make any explicit findings of fact regarding whether LTC accepted the distributor's offers. However, to the extent that the court’s legal conclusion that a sale occurred implicitly finds that LTC accepted the purchase order, then that finding was clearly erroneous. As described infra, there is no evidence of record showing an objective manifestation of assent by LTC to the distributors, and without such evidence there can be no finding of an acceptance.
. We note that there appears to be some confusion over whether evidentiary rulings in patent cases should be reviewed under the law of the regional circuit or under Federal Circuit law.
Compare Applied Med. Res. Corp.,