FOR PUBLICATION
 UNITED STATES COURT OF APPEALS
      FOR THE NINTH CIRCUIT

MEDTRONIC, INC.,                       
                          Plaintiff,
               and
LOS ANGELES BIOMEDICAL
RESEARCH INSTITUTE AT HARBOR-
UCLA MEDICAL CENTER,                         No. 06-16229
             Intervenor-Appellant,
               v.                             D.C. No.
                                           CV-04-02201-JSW
GEOFFREY WHITE,                               OPINION
             Defendant-Appellee,
               v.
EDWARDS LIFESCIENCES LLC;
ENDOGED RESEARCH PTY LIMITED,
           Third-Party-Plaintiffs.
                                       
       Appeal from the United States District Court
          for the Northern District of California
        Jeffrey S. White, District Judge, Presiding

                  Argued and Submitted
        April 15, 2008—San Francisco, California

                     Filed May 15, 2008

     Before: Warren J. Ferguson, Stephen S. Trott, and
             Sidney R. Thomas, Circuit Judges.

                   Opinion by Judge Trott



                           5541
5544    LOS ANGELES BIOMED. RESEARCH INST. v. WHITE


                        COUNSEL

Linda F. Callison and Gordon C. Atkinson, Cooley Godward
LLP, Palo Alto, California, for the intervenor-appellant.

Mark E. Haddad, Sidley Austin LLP, Los Angeles, California,
for the defendant-appellee.


                        OPINION

TROTT, Circuit Judge:

   Los Angeles Biomedical Research Institute at Harbor-
UCLA Medical Center (“L.A. Biomed”) appeals the entry of
judgment resulting from a jury verdict in favor of defendant
Dr. Geoffrey White in a contract dispute over ownership of a
patent. L.A. Biomed alleges that the jury instructions con-
tained a number of prejudicial errors. We have jurisdiction
under 28 U.S.C. § 1291, and we reverse and remand for a new
trial.
        LOS ANGELES BIOMED. RESEARCH INST. v. WHITE               5545
                                I

                       BACKGROUND

Factual Background

   L.A. Biomed is a non-profit medical research institute that
allows visiting researchers to use its state-of-the-art equip-
ment and facilities. Researchers wishing to use the facilities
are generally required to sign a Patent and Copyright Agree-
ment (“P&C Agreement”). In 1985, Dr. White became an
assistant professor of surgery at Harbor-UCLA Medical Cen-
ter and signed L.A. Biomed’s P&C Agreement. The P&C
Agreement states in pertinent part:

       This agreement is made by me with . . . Harbor-
    UCLA Medical Center, a non-profit corporation,
    hereinafter referred to as the “Institute”, in part con-
    sideration of my employment . . . and/or my utiliza-
    tion of Institute research facilities.

       I understand and agree that every possibly patent-
    able device, process, or product hereinafter referred
    to as “invention”, which I conceive and/or reduce to
    practice while employed by the Institute, or during
    the course of my utilization of any Institute research
    facilities, shall be examined by the Institute to deter-
    mine rights and equities therein in accordance with
    the Institute’s Patent and Copyright Policy.

    ....

       I further agree that, in the event any such inven-
    tion and/or work shall be deemed by the Institute to
    be patentable . . . and the Institute desires . . . to seek
    patent . . . protection therein, I shall execute any doc-
    uments and do all things necessary . . . to assign to
    the Institute all rights, title and interest therein and
5546     LOS ANGELES BIOMED. RESEARCH INST. v. WHITE
    to assist the Institute in securing patent . . . protec-
    tion therein. . . .

(emphasis added).

   “Conceive” and “reduce to practice” as used in the P&C
Agreement are terms of art used in patent and inventorship
law. Because these terms are of central importance to under-
standing the facts and analysis of this case, we define them at
the outset.

   “Conceive” means, “the formation in the mind of the inven-
tor, of a definite and permanent idea of the complete and
operative invention.” Burroughs Wellcome Co. v. Barr Labs.,
Inc., 

, 1228 (Fed. Cir. 1994) (internal quotation
marks omitted). “Conception is complete only when the idea
is so clearly defined in the inventor’s mind that only ordinary
skill would be necessary to reduce the invention to practice,
without extensive research or experimentation.” 

to establish . . . reduction to practice, the inventor
must prove that: (1) he constructed [the invention]; and (2) he
determined that the invention would work for its intended
purpose.” Cooper v. Goldfarb, 

, 1327 (Fed. Cir.
1998).

   Dr. White continued to work at Harbor-UCLA Medical
Center until 1989, when he returned to Australia where he is
a native and citizen. There, he began exploring endovascular
repair of abdominal aortic aneurysms. An aortic aneurysm is
the “ballooning” of the aorta that if untreated typically results
in rupture and death. Endovascular surgery repairs such aneu-
rysms without invasive surgery by delivering a fabric tube
referred to as a graft from an incision in the upper thigh
through the femoral artery to the location of the aneurysm,
where the graft essentially replaces the ballooning section of
the aorta.
         LOS ANGELES BIOMED. RESEARCH INST. v. WHITE       5547
   By 1992, Dr. White, and his colleague Dr.Weiyun Yu, had
begun to develop—and eventually invented—a new device
that would make it easier precisely to place a graft into the
aorta. The device, referred to generally as a graft attachment
device (“GAD”), allows a surgeon to adjust the length of a
graft within the aorta by placing one graft inside another and
then simply adjusting the overlap between them.

   At trial, and now on appeal, the parties hotly disputed when
the patented GAD was conceived and reduced to practice by
Drs. White and Yu. What seems to be agreed on is that the
doctors worked on the development of a GAD in Australia
until October 1992 when they first attempted, unsuccessfully,
to place a version of it into a patient in Sydney. Dr. White
then arranged for him and Dr. Yu to have access to L.A.
Biomed’s facilities during December 1992 and January 1993
to do certain bench tests and to build and implant miniature
versions of the GAD into dogs. Dr. Yu was not asked to and
did not sign a P&C Agreement at that or any other time. This
fact became part of Dr. White’s defense.

   Besides performing the aforementioned tests and experi-
ments, during those two months Dr. White also did other test-
ing of the GAD at the VA Long Beach and UC Irvine
Hospitals. The two doctors then returned to Australia and con-
tinued to work on the GAD’s development until October,
1993, when Dr. White returned to the VA Long Beach Hospi-
tal and successfully implanted an overlapping GAD-graft into
a human.

   Without notice to L.A. Biomed, Drs. White and Yu filed
for two patents on the GAD in 1998. The patents were issued
to them in 2003. See U.S. Patent No. 6,582,458 (filed May 1,
1998) (issued June 24, 2003); U.S. Patent No. 6,613,073
(filed Aug. 11, 1998) (issued Sept. 2, 2003). These patents
include a number of drawings illustrating the distinct features
of the GAD. See, e.g., ‘458 Patent figs.1-6. As part of the pro-
cess for obtaining these patents, Dr. White filed in 2000 a
5548      LOS ANGELES BIOMED. RESEARCH INST. v. WHITE
sworn declaration with the U.S. Patent and Trademark Office
(“the PTO declaration”), which described the progression of
the GAD’s development. It stated in pertinent part:

        (7) Between 1989 and 1992 I began to develop,
      with Dr. Weiyun Yu, the [GAD].

         (8) In 1992 I came to the United States [and] dis-
      closed to Dr. [Samuel E.] Wilson1 the types of mate-
      rials that could be used to make the graft, how the
      graft was to be assembled and how the aspects of the
      graft design would function together. . . . I believe
      that the disclosure made to Dr. Wilson was of suffi-
      cient detail that [it] would have enabled one of ordi-
      nary skill in the art to make an endovascular graft
      having such features.

         (9) In December of 1992 I undertook bench test-
      ing of an endovascular graft having features dis-
      closed in [the patent application]. This bench testing
      occurred at [L.A. Biomed]. The bench testing
      included placing an endovascular graft inside
      another graft and balloon expanding the endovascu-
      lar graft therein. The grafts used were made by Dr.
      Weiyun Yu and me, at my instruction, and had the
      features of a wireform supported prosthesis which
      could be overlapped with another similar prosthesis.
      More than twenty grafts were tested and the bench
      tests indicated that one graft could be supported
      within another.
  1
    Dr. Wilson testified about the discussions he had with Dr. White in the
early 1990s about the GAD’s development. Dr. Wilson was Chief of Sur-
gery at Harbor-UCLA Medical Center in the 1980s when Dr. White
worked there. The two doctors apparently maintained a personal friend-
ship and professional confidence. In 1992, Dr. Wilson moved his practice
to the University of California, Irvine.
         LOS ANGELES BIOMED. RESEARCH INST. v. WHITE         5549
       (10) . . . . In September of 1993 . . . I performed
    a procedure in Sydney, Australia, wherein I repaired
    a patient’s aortic aneurysm using an endovascular
    graft. . . .

       (11) . . . . On October 6, 1993, I performed an
    endovascular repair of an abdominal aortic aneurysm
    utilizing an endovascular graft having the features of
    a wireform supported prosthesis which could be
    overlapped within another prosthesis. . . .

Procedural Background

   This suit arose out of patent infringement litigation that
began after the patents issued. During the infringement litiga-
tion, L.A. Biomed intervened in the case when it learned of
Dr. White’s PTO declaration claiming conclusive testing at its
facilities. The district court stayed that litigation until the
ownership issues were resolved.

   Thus, the trial from which this appeal arises became a con-
tract dispute over the ownership of the GAD patents. Specifi-
cally, the issues were whether the P&C Agreement continued
to apply to Dr. White in 1992 and 1993 and, if so, whether he
had conceived and/or reduced to practice the patented GAD
while at L.A. Biomed. At issue also were the legal implica-
tions of Dr. Yu’s involvement in the process.

   L.A. Biomed argued that during December 1992 and Janu-
ary 1993, while at L.A. Biomed, Drs. White and Yu con-
ceived, built, and successfully tested a “Z-ring GAD device”
that was materially different from an earlier “O-ring GAD
design” developed in Australia. L.A. Biomed claimed that,
while using its facilities, Dr. White conceived and/or reduced
to practice the patented version of the GAD by inventing and
testing four new, distinct features that were central to the ulti-
mate operative invention, which L.A. Biomed refers to as the
“Z-ring GAD device.” Accordingly, L.A. Biomed argued that
5550      LOS ANGELES BIOMED. RESEARCH INST. v. WHITE
Dr. White was obligated by the P&C Agreement to assign his
patent rights to the GAD to L.A. Biomed.

   Dr. White, on the other hand, argued that he had no obliga-
tion under the P&C Agreement, even if it still applied to him,
because by October of 1992, he had already conceived, in
Australia, the “Z-ring GAD device” with the same features
that L.A. Biomed contended were conceived and tested at
L.A. Biomed. Specifically, he argued that while at L.A.
Biomed, he and Dr. Yu only rebuilt identical or miniature ver-
sions of the GAD already conceived in Australia and repeated
tests they had already performed in Australia. He argued also
that the GAD was reduced to practice when he successfully
implanted it into a patient at the VA Long Beach Hospital in
October of 1993.

   Both parties submitted numerous versions of their proposed
jury instructions. After hearing the parties on the proposed
instructions, the district court issued its final jury instructions.
The relevant final instructions there, and now at issue on
appeal, are those entitled “Work of Dr. Yu,” “Conceive,” and
“Reduce to Practice”2 (cumulatively, “the disputed instruc-
tions”).

   At the conclusion of the trial, the jury returned a special
verdict in favor of Dr. White. Although it found that the P&C
Agreement existed between L.A. Biomed and Dr. White at all
relevant times, the jury found also that L.A. Biomed had
failed to prove that Dr. White had conceived or reduced to
practice the GAD during the course of his utilization of L.A.
Biomed research facilities. The district court entered judg-
ment on that verdict.
  2
    While the proposed instruction was titled “Reduction to Practice,” the
instruction given to the jury was titled “Reduce to Practice.” For the sake
of clarity and consistency, we refer to the proposed and the final versions
of the instruction generically as the "Reduce to Practice" instruction.
         LOS ANGELES BIOMED. RESEARCH INST. v. WHITE       5551
                               II

                        DISCUSSION

A.   Standard of Review.

   “The standard of review on appeal for an alleged error in
jury instructions depends on the nature of the claimed error.”
Dang v. Cross, 

, 804 (9th Cir. 2005) (internal
quotation marks omitted). “We review de novo whether the
instructions misstated the law.” 

marks
omitted). “We review a district court’s formulation of jury
instructions in a civil case for abuse of discretion.” 

is de novo because L.A. Biomed alleges misstatements
of law.

   Where an error in instructing the jury is found, prejudice is
presumed and “the burden shifts to the [appellee] to demon-
strate that it is more probable than not that the jury would
have reached the same verdict had it been properly instruct-
ed.” Galdamez v. Potter, 

, 1025 (9th Cir. 2005)
(internal quotation marks omitted).

B.   The Application of Agency Law to the Exclusion of
     Patent Law in the Disputed Instructions was
     Prejudicial Error.

   At trial, evidence and testimony pertaining to Dr. Yu’s
efforts in developing the GAD created questions about
whether and to what extent his work at L.A. Biomed’s facili-
ties could be used to show a breach of the P&C Agreement
by Dr. White. Part of Dr. White’s defense was that absent evi-
dence of agency between Dr. White and Dr. Yu, L.A. Biomed
had no right to claim any rights in the patented devices that
they later conceived or reduced to practice at L.A. Biomed.
L.A. Biomed proposed that no agency instruction be given
and that the “Conceive” instruction should contain a para-
graph summarizing the patent law principle of co-
5552        LOS ANGELES BIOMED. RESEARCH INST. v. WHITE
inventorship—specifically, that an “invention can be con-
ceived jointly” where each inventor “makes a significant con-
tribution to [its] conception.”3 L.A. Biomed proposed further
that the “Reduce to Practice” instruction include a paragraph
summarizing the similar patent law principle that “[a]cts
related to reduction to practice that are performed by a co-
inventor . . . should be considered as if they had been per-
formed by the inventor himself.”4 L.A. Biomed stated also in
its proposed instructions that, in the alternative, if any agency
instruction was given, a model agency instruction should be
given. Dr. White proposed a long instruction that applied
agency law.
  3
   The proposed “Conceive” instruction was accompanied by citations to
a number of Federal Circuit decisions laying out the applicable principles
of patent and inventorship law, including, for example, 

, and Fina Oil and Chem. Co. v. Ewen, 

,
1473 (Fed. Cir. 1997). The proposed instruction read:
        The contract between LA Biomed and Dr. White uses the term
      “conceived.” . . . . [Defines “conceive” as it is used in patent law]
         An invention may be conceived jointly. To be a joint inventor,
      one must make a significant contribution to the conception of the
      invention. Persons may be inventors even though they do not
      physically work together or make the same type or amount of
      contribution, or contribute to the subject matter of each feature of
      the invention.
   4
     The proposed “Reduce to Practice” instruction was accompanied by
citations to a number of Federal Circuit decisions laying out the applicable
principles of patent and inventorship law, including, for example, 

, 1331. The relevant part of the proposed instruction read:
         The contract between LA Biomed and Dr. White uses the term
      “reduced to practice.” . . . . [Defines “reduction to practice” as it
      is used in patent law]
      . . . . [Discusses amount of testing necessary, according to patent
      law, for “reduction to practice” to have occurred]
         An inventor need not personally reduce to practice his inven-
      tion. Acts related to reduction to practice that are performed by
      a co-inventor or a person working with or for the inventor should
      be considered as if they had been performed by the inventor him-
      self.
            LOS ANGELES BIOMED. RESEARCH INST. v. WHITE                5553
   The district court eventually decided to give the model
agency instruction—entitling it “Work of Dr. Yu”5 —and not
to include the proposed co-inventor language in the “Con-
ceive” and “Reduce to Practice” instructions. The district
court opined at the charge conference that the model agency
instruction was the best method for dealing with Dr. Yu’s
contributions to the invention of the GAD because it “seems
to capture the points that both sides want to argue without
unduly restricting the jury and giving them some structure.”

  1.     L.A. Biomed Preserved Its Objections to the
         Disputed Instructions.

   As an initial matter, Dr. White argues that L.A. Biomed
waived any objection to the disputed instructions either by
failing to object at the charge conference or by later acquiesc-
ing to the district court’s final instructions.

  [1] For an objection to a jury instruction to be valid, the
objection must be made “on the record, stating distinctly the
  5
   The model agency instruction given to the jury read:
                              Work of Dr. Yu
         You have heard testimony about both Dr. Geoffrey White and
      Dr. Weiyun Yu. Dr. Yu is neither a party to this case nor a signa-
      tory to the Patent and Copyright Agreement at issue. L.A.
      Biomed makes no separate claim against Dr. Yu in this case.
         L.A. Biomed claims that Dr. Yu was Dr. White’s agent and
      that Dr. White is therefor responsible for Dr. Yu’s conduct with
      respect to the Patent and Copyright Agreement.
         If L.A. Biomed proves that Dr. White gave Dr. Yu authority
      to act on his behalf, then Dr. Yu was Dr. White’s agent. This
      authority may be shown by words or may be implied by the par-
      ties’ conduct. This authority cannot be shown by the words of Dr.
      Yu alone.
        If L.A. Biomed does not prove that Dr. Yu was Dr. White’s
      agent, you should not consider Dr. Yu’s conduct in reaching your
      decision.
5554     LOS ANGELES BIOMED. RESEARCH INST. v. WHITE
matter objected to and the grounds for the objection.” FED. R.
CIV. P. 51(c)(1). In its proposed jury instructions L.A. Biomed
stated clearly its position that “no agency instruction should
be given as an agency instruction is irrelevant and confusing.
L.A. Biomed submits that the effect of [Drs. Yu and White’s]
collaboration is more appropriately addressed in L.A.
Biomed’s proposed conception and reduction to practice
instruction[s].” At the charge conference, L.A. Biomed’s
counsel distinctly stated her objection to the application of
agency law, instead of patent law, in both the “Work of Dr.
Yu” and “Reduce to Practice” instructions. Thus, L.A.
Biomed preserved for appeal its objection to these two
instructions.

   Dr. White argues also that L.A. Biomed waived its objec-
tion by acquiescing to the “Conceive” instruction when its
counsel responded at the charge conference, “I think that’s
fine,” after the district court read the version that was ulti-
mately presented to the jury. Although “[t]his court has
enjoyed a reputation as the strictest enforcer of Rule 51,” we
recognize a limited exception “[w]here the district court is
aware of a party’s concerns and further objection would be
unavailing.” Glover v. BIC Corp., 

, 1326 (9th Cir.
1993) (internal quotation marks omitted). The exception is
available “when (1) throughout the trial the party argued the
disputed matter with the court, (2) it is clear from the record
that the court knew the party’s grounds for disagreement with
the instruction, and (3) the party offered an alternative instruc-
tion.” 

marks omitted).

   [2] At trial, after L.A. Biomed submitted its proposed
instructions but prior to L.A. Biomed’s alleged acquiescence,
the district court had rejected the co-inventorship language
from the “Conceive” instruction. Specifically, the court stated,
“I was concerned when I was reading some of the proposed
. . . instructions on this issue that we were kind of slopping
over into . . . inventorship, co-inventorship, which is a totally
different body of law that this jury is not going to be consider-
           LOS ANGELES BIOMED. RESEARCH INST. v. WHITE                 5555
ing.” Thus it is clear that the district court was aware of L.A.
Biomed’s disagreement with the “Conceive” instruction and
that further objection at the charge conference would have
been unavailing.6 Therefore, we conclude that L.A. Biomed’s
objection to the “Conceive” instruction was preserved. See 

Disputed Instructions Were Given In Error
        Because the P&C Agreement Required the
        Application of Patent Law.

   [3] L.A. Biomed argues that the district court erred by not
including the co-inventorship language in the jury instructions
because the dispute is over the ownership of a patent and is
rooted in a contract that dictates that patent and inventorship
law should control its interpretation. L.A. Biomed correctly
points out that under the settled law of inventorship, each co-
inventor who makes a significant contribution to an invention
owns an undivided interest in the corresponding patent. See,
e.g., Ethicon, Inc. v. U.S. Surgical Corp., 

, 1465
(Fed. Cir. 1998); Fina 

. L.A. Biomed
thus argues that the agency instruction was given in error
because the issue was not whether Dr. Yu acted as Dr.
White’s agent, but whether Dr. White made a substantial con-
tribution to the conception of an invention while using its
facilities.

   Dr. White, on the other hand, argues that giving an instruc-
tion applying agency law was perfectly appropriate in this
breach of contract case where the contract at issue pertains
only to inventions Dr. White conceived and/or reduced to
practice while utilizing L.A. Biomed’s facilities.
  6
    It is also worth noting that at the beginning of the charge conference,
and prior to the alleged acquiescence, the district court stated, “any
instruction that [was] requested that I don’t give [is] deemed . . . pre-
served, and you don’t need to reassert those. So my order is that . . . those
objections and requests will be preserved just by virtue of having
requested those instructions.”
5556     LOS ANGELES BIOMED. RESEARCH INST. v. WHITE
   [4] We first acknowledge that “issues of patent ownership
are distinct from questions of inventorship.” Israel Bio-Eng’g
Project v. Amgen, Inc., 

, 1263 (Fed. Cir. 2007)
(hereinafter “IBEP”). Additionally, we recognize that the Cal-
ifornia Court of Appeal has held that “the trial court correctly
applied contract principles in resolving [a] dispute over [a]
patent agreement” nearly identical to the P&C Agreement.
Shaw v. Regents of Univ. of Cal., 

, 854
(Cal. Ct. App. 1997).

   In Shaw, a professor sued the University of California when
it instituted a new patent policy that reduced the percentage
of net royalties payable to him under the patent agreement he
signed when he began his employment. 

One of the
central issues was whether the district court erred in applying
contract law “rather than the standard of review for a manda-
mus action.” 

California’s Court of Appeal found
that contract law applied and looked to the language of the
contract to determine “the meaning and effect of the patent
agreement.” 

   Thus, we turn to the language of the P&C Agreement to
determine whether the parties intended for the rules of inven-
torship or the rules of agency to apply.

   [5] First, we note that this contract employed a number of
terms that reflect well settled principles of patent and inven-
torship law such as “conceive” and “reduce to practice.”
Additionally, L.A. Biomed’s P&C Policy reflects principles
of patent and inventorship law. Parties may incorporate the
terms of other documents into a contract “so long as [the con-
tract] guides the reader to the incorporated document.” 

(internal quotation marks and alteration omitted). The
P&C Agreement states that “every possibly patentable device
. . . . shall be examined by L.A. Biomed to determine rights
and equities therein in accordance with L.A. Biomed’s Patent
and Copyright Policy.” This is sufficient to incorporate the
P&C Policy into the P&C Agreement. See 

BIOMED. RESEARCH INST. v. WHITE         5557
   [6] L.A. Biomed’s P&C Policy states: “In its consideration
of matters relating to each particular patent . . . case or situa-
tion, [LAB’s] Patent Board shall take into consideration the
principles laid down in the patent . . . laws and in the court
decisions of the United States.” This strongly indicates that
the parties intended for patent law to be applied when deter-
mining patent ownership rights and equities.

   [7] Therefore, we conclude that the pertinent language of
the contract indicates convincingly that the parties intended
for patent law to apply in interpreting the P&C Agreement.
Consequently, we conclude that it was clear error for the dis-
trict court to give the agency instruction and to exclude the
co-inventorship language proposed by L.A. Biomed. Agency
was a red herring.

   Because we find that the disputed instructions misstated the
law, we presume prejudice and the burden shifts to Dr. White
“to demonstrate that it is more probable than not that the jury
would have reached the same verdict had it been properly
instructed.” 

(internal quotation
marks omitted). We conclude that Dr. White did not meet this
burden and that the error was prejudicial because it allowed
the jury to decide the case on a legally impermissible ground
—specifically, a reasonable juror could have found that Drs.
White and Yu conceived or reduced to practice the patented
GAD at L.A. Biomed but still would have been compelled
nonetheless to return a verdict for Dr. White if they found, as
Dr. White argued, that Dr. Yu was not acting as his agent. See
Heller v. EBB Auto Co., 

, 1441 (9th Cir. 1993)
(reversing and remanding a jury verdict where the jury could
have based its verdict on a finding inconsistent with the law
due to the district court’s erroneous instruction).

   Dr. White now contends that Dr. Yu’s role was a side note
to the centerpiece of his defense that the GAD was conceived
in Australia and reduced to practice at the VA Long Beach
Hospital. This assertion and the record, however, are insuffi-
5558      LOS ANGELES BIOMED. RESEARCH INST. v. WHITE
cient to carry Dr. White’s burden. For example, at closing Dr.
White’s counsel argued to the jury that Dr. Yu’s efforts were
independent of Dr. White’s. He argued, “What did Dr. Yu tell
you when he sat in the stand? These were his designs. His
designs. He builds the grafts. He designed the wireforms.” He
went on to argue that Dr. Yu “clearly was working on his
own. He clearly designed these devices.” This is enough to
show that the instructional error was not harmless. See Gizoni
v. Sw. Marine Inc., 

, 1141-42 (9th Cir. 1995)
(finding error in jury instruction prejudicial “in light of
[appellee’s] closing arguments”).

  Because the agency instruction combined with the absence
of a recitation of settled co-inventorship law was error, and
because Dr. White has failed to demonstrate that the error was
more probably than not harmless, we reverse the district
court’s judgment and remand this case to the district court for
a new trial consistent with this opinion.

C.     The “Corroboration Instruction” was Prejudicial
       Error.

   L.A. Biomed alleges that the final paragraph of the “Re-
duce To Practice” instruction (“the corroboration instruction”)
erroneously misstated the law because it failed to inform the
jury that any admission against interest by Dr. White need not
be corroborated by independent evidence. L.A. Biomed is
correct. The first sentence of the corroboration instruction
summarizes an accepted principle of patent and inventorship
law: “To prove reduction to practice by testimony from a per-
son who claims to have reduced to practice a device, that tes-
timony must be corroborated by independent evidence.” See,
e.g., 

. The instruction goes on to
explain what is sufficient corroborating evidence. L.A.
Biomed argues not that the corroboration instruction itself
misstates the law, but that the district court misstated the law
by failing to acknowledge that an inventor’s admissions-
against-interest need not be corroborated.
         LOS ANGELES BIOMED. RESEARCH INST. v. WHITE          5559
   [8] We conclude that the corroboration instruction was
given in error. The Federal Circuit’s predecessor court has
held that in a dispute over the date of conception of an inven-
tion, “all that is necessary to constitute an admission is a pre-
vious statement by an adversary party which is inconsistent
with the position he is taking in litigation.” Technitrol, Inc. v.
United States, 

, 1370 (Ct. Cl. 1971). In other
words, although corroborating evidence is required of an
inventor pursuing a patent to prove the date of conception or
reduction to practice of an invention, such corroborating evi-
dence is not required when offered by an adversary party
against an inventor as an admission-against interest.

   [9] Thus, we conclude that the corroboration instruction,
without a corresponding admission against interest instruc-
tion, was given in error because it misstated the law by requir-
ing corroborating evidence.7

   Furthermore, we conclude that this instruction was prejudi-
cial. In the PTO declaration, Dr. White stated that he per-
formed bench testing at L.A. Biomed on a version of the
GAD that “had the features of a wireform supported prosthe-
sis which could be overlapped with another similar
prosthesis”—one of the features L.A. Biomed argued was
central to the invention of the “Z-ring GAD device.” He
stated further in the PTO declaration that “the bench tests
indicated that one graft could be supported within another.”

   [10] Dr. White argues that this was not an admission
against interest and thus could not be prejudicial because the
PTO declaration was consistent with his position at trial. We
disagree, however, and conclude that a reasonable juror could
have found that this statement from the PTO declaration con-
stituted at least some convincing evidence that Dr. White con-
ceived and/or reduced to practice the patented GAD at L.A.
  7
   We have considered Dr. White’s other arguments on this issue and
determined that they are without merit.
5560      LOS ANGELES BIOMED. RESEARCH INST. v. WHITE
Biomed. Yet, the corroboration instruction inappropriately
prohibited the jury from making such a finding unless there
was also sufficient independent corroborating evidence sup-
porting it. Thus, because a reasonable juror could have based
a verdict in favor of L.A. Biomed on Dr. White’s declaration
in the absence of the corroboration instruction, we conclude
that it was prejudicial and a new trial is necessary for this rea-
son as well. See 

.

D.     The Identification of the Invention.

   L.A. Biomed complains that the district court improperly
blurred an important distinction between Dr. White’s “O-ring
GAD design” and the successful “Z-ring GAD device” when
the court held that L.A. Biomed’s contract claim would be
limited to whether Dr. White invented an “overlapping GAD-
graft device” at L.A. Biomed. Given the exact terms of the
agreement Dr. White signed, which required him to disclose
“every possible patentable device,” this contention has merit.
It is clear that more precise jury instructions on this central
point of contention needed to have been given.

   We leave the rectification of this issue to the district court
on remand. However, given the parties’ manifest willingness
endlessly to dispute on appeal who argued what and when,
and who presented what in the multiple conferences on
instructions, we would advise both the district court and the
parties—now that they will have a fresh start—to take great
care to respect Rule 51 and to leave nothing either to infer-
ence or to the imagination. We recognize the convoluted his-
tory of this case in the trial court, but it turns out that allowing
the parties to “deem preserved” previous objections created an
unnecessary Rule 51 battlefield on appeal.

                                III

                        CONCLUSION

  Because we conclude that both the agency instruction and
the corroboration instruction were given in error and that each
       LOS ANGELES BIOMED. RESEARCH INST. v. WHITE   5561
was prejudicial, we REVERSE and REMAND to the district
court for a new trial consistent with this opinion.

  REVERSED and REMANDED.