639 S.W.2d 726 (1982)
The FIRESTONE TIRE & RUBBER COMPANY, Appellant,
v.
Artie LITTLE, Appellee.
No. 82-17.
Supreme Court of Arkansas.
October 4, 1982.
Black well, Sanders, Matheny, Weary & Lombardi, Kansas City, Mo., and Rose Law Firm, P.A., Little Rock, for appellant.
Niewald, Risjord & Waldeck, Kansas City, Mo., and Shackleford, Shackleford & Phillips, P.A., El Dorado, for appellee.
HICKMAN, Justice.
This is a products liability case. In 1978, Artie Little, age 82, was walking by the roadside in Strong, Arkansas, when she was hit with a rim that came off the wheel of a passing truck. She filed suit for damages against the owner of the truck and trailer, Harvey Shelton, the owner of a service station who fixed a flat on the wheel of the trailer that day, Jackson Smith, and Firestone, the manufacturer of the rim. At trial, the jury exonerated both the owner of the truck and the service station owner, and awarded Artie Little $150,000 compensatory damages, and $200 punitive damages against Firestone.
The judgment has to be reversed because Firestone asked the day before trial whether the plaintiff and the two other defendants, Shelton and Smith, had entered into a "Mary Carter Agreement." A Mary Carter agreement is one in which a plaintiff secretly agrees with a defendant that if the plaintiff recovers from another defendant, the agreeing defendant's liability will be reduced. Those agreements were so named when one arose in Florida in Booth v. Mary Carter Paint Company, 202 So.2d 8 (Fla. App. 2d 1967). Firestone's counsel asked before trial if any agreements had been made whereby Shelton's or Smith's liability would be reduced if Artie Little recovered against Firestone. Smith's and Shelton's attorneys objected and the trial court did not order the disclosure of any such agreement. On appeal, the appellee argues that Firestone knew or suspected such an agreement existed several months before trial arid should have attempted to discover that information far in advance of the day before trial. We do not find Firestone waived their right to object.
There seems to be little doubt of the existence of some sort of agreement and we hold that the trial court was wrong in not requiring the agreement to be disclosed. Furthermore, we join those states that hold such an agreement is not only discoverable but also may be admitted into evidence.
The state courts that have considered this question are split to some degree on whether such an agreement is unethical or against public policy. See Lum v. Stinnett, 87 Nev. 402, 488 P.2d 347 (1971), and Lubbock Manufacturing Co. v. Perez, 591 S.W.2d 907 (Tex.Civ.App.1979). But we have no hesitation in joining those that require a full disclosure in cases such as this. General Motors Corp. v. Lahocki, 286 M.D. 714, 410 A.2d 1039 (Md.App.1980); Gatto v. Walgreen Drug Co., 61 Ill.2d 513, 337 N.E.2d 23 (Ill.1975); Ward v. Ochoa, 284 So.2d 385 (Fla.1973); Pellett v. Sonotone Corp., 26 Cal.2d 705, 160 P.2d 783 (1945); See 65 A.L.R.3d 602. The testimony of Shelton and Smith was critical to Artie Little's case against Firestone. And, as it turns out, their testimony was no doubt a strong factor in the jury's determination that Firestone's RH5° rim was the sole cause of the accident and the complete exoneration of Shelton and Smith. It is readily apparent why the jury should know of any deals these parties made. As the Florida court said in the case of Ward v. Ochoa, supra:
Secrecy is the essence of such an arrangement, because the court or jury as trier of the facts, if apprised of this, would likely weigh differently the testimony and conduct of the signing defendant as related to the non-signing defendants. By painting a gruesome testimonial picture of the other defendant's conduct or, in some cases, by admissions against himself and the other defendants, he could diminish or eliminate his own liability by use of the secret `Mary Carter Agreement.'
The search for the truth, in order to give justice to the litigants, is the primary duty of the courts. Secret agreements between plaintiffs and one or more of several multiple defendants can tend to mislead judges and juries, and border on collusion.
Firestone's most ardent argument is that the case should be dismissed because it is impossible that the wheel in evidence is the accident wheel.[1] The entire lawsuit focuses on the RH5° rim base which Firestone had manufactured from 1946 to 1973. It is not disputed that almost twenty-five million such rims were manufactured. The rim base consists of two parts, an outer ring, which is the part that supposedly flew off and struck Artie Little, and the base itself, which is the widest part of the rim and on which a disc is either bolted or welded. The disc is the part that contains the bolt holes.
Firestone's argument that the wheel in question could not have caused the accident is premised on the testimony of the truck driver, Shelton, and Smith, the wheel introduced into evidence, testimony of certain Firestone employees, and other exhibits. The testimony was: The owner, Shelton, said he bought the trailer in question in 1968 and had never changed the wheels. The previous owner testified there had never been a multi-piece wheel on the trailer. (That would exclude the wheel in question.) Baker, the driver, said the day before the accident he noticed the left front outside wheel on the tandem trailer was flat. The next morning Jackson Smith fixed the flat with Baker's help. Baker left Smith's station and drove about four blocks when he saw Artie Little walking along and then saw her disappear. He stopped his truck and ran to her assistance. He found that the tire that had been fixed had exploded and Artie Little was thrown into a ditch. Apparently the rim had come off and struck her. The base and tire were still on the truck, but the tire was in shreds. What happened next to the outer rim and wheel is somewhat in dispute.
Smith said the ring and rim were brought to his station where he kept them inside the building on a junk pile until someone picked them up. The truck owner, Shelton, said he picked them up at the accident scene, took them home and cut out the "eye" of the disc so it could not be used by anyone else. The "eye" is merely the center of the disc where the bolt holes are located. Baker, the driver, said he also went to Shelton's. Shelton said he had both the remaining disc and the ring which were picked up at his place by one of the lawyers. The "eye" he cut out was never produced. It was this wheel assembly that was introduced by Artie Little as the Firestone product that caused the accident. Shelton testified that he was convinced that the wheel assembly in evidence was the one that came off the truck; Baker testified that the wheel in evidence was exactly like the accident wheel.
Firestone produced evidence that the wheel in evidence had only "five hand" holes and they never produced a "five hand" hole assembly that only had six bolt holesit either had eight or ten bolt holes. A technical advisor for Firestone examined the trailer and testified that the axle on Shelton's trail could only take a six bolt wheel. Furthermore, when the assembly was shown to Jackson Smith at trial, he said that the assembly could not have been the one he put on that day because the outer ring would easily slip off the base and it would not hold as an RH5° should. He said, however, that he was not mistaken that the wheel he put on the truck that day was a RH5° he had handled thousands of them. Furthermore, he said he knew they were dangerous and had cautioned his customers that used them.
Mr. Roger B. McCarthy, an engineer from California employed by Failure Analysis Associates, said he tried to duplicate the rim assembly in evidence. He said such an assembly would not hold sufficient pressure to be driven (which would be 85 to 90 pounds), and would come apart at about 45 pounds pressure. He said he filled the tire three times with water and it came apart each time at about 45 pounds.
Firestone's argument, of course, is one that should be made to the jury. We cannot find that it was physically impossible that the wheel actually introduced was not the accident wheel. General Motors Corp. v. Tate, 257 Ark. 347, 516 S.W.2d 602 (1974), is cited by the appellant as controlling on the issue of the weight to be given the defendant's evidence of the physical impossibility of the accident occurring in the way the plaintiff claims. We held in General Motors that such evidence may not be arbitrarily and capriciously disregarded by the trier of fact. In this case the defendant's facts were disputed. The appellee produced witnesses that said the wheel in evidence was the accident wheel. While Firestone raised serious questions about the authenticity of this rim, it was for the jury to decide which witnesses to believe. Circle Realty Co. v. Gottlieb, 267 Ark. 160, 589 S.W.2d 574 (1979). Actually we cannot say it was essential in this particular case that Artie Little produce the accident wheel.
There is no doubt Firestone was harmed by the admission of certain documentary evidence, and its admissibility was the subject of most of the other objections raised by Firestone. One of the most damaging documents was a letter written in 1969 to an attorney involved in litigation over a multi-piece wheel. It was written by Paul Hykes, an engineer who worked for the Budd Company, which bought the Firestone RH5° rims and used them in the manufacture of wheels. The letter said, in part:
I am convinced that the RH5° Rim, aside from the safety aspect, has more than its share of field failures.
The RH5° Rim is a clever design, two continuous rings vs. one split ring on other two piece demountable side ring rim and designs or one split lock ring on the three piece versions. This permits lighter weight, lower cost. The material used is comparable to that used in other designs of rims. Its faults are:
1. It is more subject to dangerous wear than many other designs.
2. The only time that this dangerous wear can be detected is when the rim is dismantled.
3. It can be put together improperly and subsequently blow-apart when new.
4. It is easier to put it together improperly when it is badly worn.
5. It is more difficult to detect improper assembly on the RH5° than it is on competing rims.
From the foregoing, you can readily understand why I ask the question, `Why has this Rim not been removed from the market?'
The letter was accepted into evidence as notice to Firestone in 1969 that its rim was deficient in several respects. Firestone argued that it was hearsay, irrelevant and not authenticated. After the trial, Firestone took Hykes' deposition and apparently he significantly qualified the statements in the letter. The letter was admitted to prove notice to Firestone, not to prove the truth of the assertions in the letter and, therefore, was admissible under Ark.Stat.Ann. § 28-1001, Rule 801 (Repl. 1979). We cannot say the trial court abused its discretion in allowing the letter; on retrial Firestone should be permitted to introduce Hykes' deposition taken after the trial.
Another objection concerning the same letter arose because W. H. Sanders, a lawyer for Firestone, wrote to other counsel for Firestone, and attached the Hykes letter. Sanders' letter was admitted into evidence; it said:
Dear Jim:
With an appalled silence I hand you herewith the report from the former Executive Engineer of the Budd Company.
I am at a loss to know why Budd personnel, or former personnel, are so critical of a product that Budd sold.
I would be most interested to know where Budd stands on this matter and whether it endorses the position taken by Morrison and Hykes.
I still can't bring myself to believe, or really give any credence to these criticisms of this rim. If these men are right, then the rim should not be sold, but it is sold, it is apparently providing good service, and from what Mr. McCusick had to say, I gather that these men could not come up with a safer type of rim to use.
The letter made its way to the Budd Company and Budd surrendered it in a lawsuit in answer to a discovery motion. Firestone argues the letter was inadmissible under Ark.Stat. § 28-1001, Rule 502, which provides a privilege for confidential attorney/client communications. We deem the privilege waived. Firestone should have never allowed the letter into the hands of Budd; by doing so Firestone has waived any right to claim the privilege. Ark.Stat. Ann. § 28-1001, Rule 510.
On rehearing, the appellant argues that a letter written by Mr. Lynn L. Bradford in October, 1979, Acting Director of the Office of Defects Investigation Enforcement of the National Highway Traffic Safety Administration, was not admissible as notice to Firestone because it was written after the accident in question. In our first opinion issued in this case we held that we could not say the trial court was wrong in finding that the letter was admissible. The arguments made by Firestone on appeal were that the letter was hearsay and Mr. Bradford did not have the legal authority to demand a recall. On rehearing, Firestone argues that the letter could not be notice because it was mailed after the accident in question. We reconsider this issue only because the matter will arise again on retrial and we should dispose of it. The letter was written after the accident in question and on the face of it we cannot say the letter referred to accidents or incidents which occurred before the accident in question, although it is very likely that is the case. On a retrial the letter should not be admitted because any probative value is outweighed by possible prejudice.
Firestone suggests throughout its brief that complaints made to Firestone are not evidence that there was a defect in the design of the wheel or that Firestone was under any duty to take any action. Subject to the trial court's discretion, evidence of the complaints is admissible as notice and are not hearsay. See McCormick's Evidence § 249 (1972 ed.). Moreover, it was up to the trial judge to decide whether these documents were relevant. See Ark.Stat. Ann. § 28-1001, Rule 104. And, it is within the province of the jury to decide if the evidence actually amounted to notice to Firestone.
Firestone argues that many other evidentiary decisions made by the trial judge were in error. Evidence of the National Highway Transportation Safety Association's investigation of the RH5° wheel, evidence that the Utah Industrial Commission issued a tentative order banning the RH5° wheel in Utah,[2] and nine memorandums, most of them interoffice between Firestone employees which discuss problems with the RH5° wheel, were admitted into evidence over the objection of Firestone. All posed questions of notice and relevancy and were subject to the trial judge's discretion. The evidence of 145 prior accidents involving the RH5° wheel admitted was also relevant to the question of notice. Arkansas Power & Light v. Johnson, 260 Ark. 237, 538 S.W.2d 541 (1976); McCormick's Evidence § 200 (1972 ed.).
Firestone argues that proffered testimony of Roger McCarthy regarding the accident rate associated with the RH5° wheel as compared to other sorts of accidents, as well as other proffered testimony, should have been admitted. McCarthy was allowed to testify extensively. The admission and exclusion of expert testimony is a matter which lies within the sound discretion of the trial judge, and we find no reversible error. White v. Mitchell, 263 Ark. 787, 568 S.W.2d 216 (1978); Phillips Construction Co. v. Williams, 254 Ark. 824, 496 S.W.3d 417 (1973).
The trial judge, in all of these evidentiary matters, must be afforded broad latitude. He, alone, has heard and seen all the evidence and he, alone, is in the best position to decide what evidence would aid the jury and what would only confuse the issues. And, unless we can say he was clearly wrong, we will not substitute our judgment for his. Ark.Stat.Ann. § 28-1001, Rule 104; See Arkansas State Highway Commission v. N.W.A. Realty Corp., 262 Ark. 440, 557 S.W.2d 620 (1977) and 3 Weinstein's Evidence ¶ 702[2] (1981).
It is unlikely that the remarks made by Artie Little's counsel regarding a former employee of Firestone's who was terminated after giving a deposition will occur on retrial. These and other remarks of counsel which were objected to are unlikely to occur during a retrial.
Firestone was precluded from admitting evidence of their long-term support of an Occupational Safety and Health Administration regulation that required employers to train their employees in the proper servicing of multi-piece wheels. There was evidence that in 1976 Firestone considered placing warning labels on the rim, but employees could not agree on what the warning should say. An expert on communications advised that a warning might cause more harm than good and the idea was dropped. Instead, Firestone launched a program to educate users how to properly service the rim. We think that the evidence should have been admitted to rebut Artie Little's allegations that Firestone's conduct was willful and wanton and for that purpose alone. See Johnson v. Niagara Machine and Tool Works, 666 F.2d 1223 (8th Cir. 1981).
On rehearing, Firestone argues that the jury's answer to one interrogatory, that Firestone did not supply the wheel in a defective condition which rendered it unreasonably dangerous, precludes that issue from being relitigated. This case was submitted on alternative theories of liability and AMI Instruction 1012 was given to the jury. The alternative theories were strict liability and negligence. The jury was asked to answer two interrogatories regarding Firestone's fault. The jury was asked:
Do you find from a preponderance of the evidence that Firestone Tire and Rubber Company manufactured a RH5° wheel assembly which was supplied in a defective condition which rendered it unreasonably dangerous, and that its defective condition was a proximate cause of the occurrence?
The jury answered no. The jury was also asked:
Do you find from a preponderance of the evidence that Firestone Tire and Rubber Company was guilty of negligence which was a proximate cause of the occurrence?
The jury's answer was yes, one hundred percent. Firestone's argument is that the jury, in effect, returned separate verdicts, and the jury's answer of "no" to the first interrogatory precludes a retrial on the issue of strict liability for a manufacturing defect because it is res judicata. Firestone cites Womack v. Brickell, 232 Ark. 385, 337 S.W.2d 655 (1960) as authority for its position. We disagree that the issue of strict liability is res judicata. We do not find that the interrogatories in this case amounted to separate verdicts. The appellee could not have appealed from the finding by the jury that Firestone was not liable under the strict liability theory. In the Womack case, either party could have appealed from the jury's finding. In a law case the verdict is an entirety which cannot be divided. Wilson v. Davis, 230 Ark. 1013, 328 S.W.2d 249 (1959).
There was substantial evidence in the record to support the jury's finding of negligence on the part of Firestone. The jury could have decided that Firestone was negligent in ignoring the notice it had received regarding the problems with the RH5° or in failing to warn of the dangerous propensities of the RH5° wheel. Since there was substantial evidence to support one of the several theories of liability, the appellant is not precluded from presenting its case on either or both theories on a retrial.
We have reviewed the record and find no other reversible errors.
Reversed and remanded.
PURTLE, J., dissents.
PURTLE, Justice, dissenting.
I would hold that the issue as to defective manufacturing of the wheel is res judicata on retrial.
NOTES
[1] Firestone concedes the wheel introduced was manufactured by Firestone. The outer ring was manufactured in 1960, the inner ring in 1971.
[2] Firestone sought to introduce the final order of the Commission which did not ban the wheel. The order was not on Firestone's pretrial list and the judge excluded it. No doubt it should be admitted on a retrial, if properly presented, to rebut the claim for punitive damages.