concurring in part and dissenting in part:
I concur in all of the majority opinion except its part IVD and except so much of its part VI(b) as holds Lumar liable for ninety percent, instead of 32/42nds, of Mrs. Simeon’s loss of consortium damages.
As to the majority’s part IVD, while I agree that the liability of Smith and Lumar should not be entirely several (so that Smith is liable only for fifty-eight percent of Simeon’s damages, and Lumar only for thirty-two percent), I am nevertheless of the opinion that traditional joint and several liability is inappropriate in light of Si-meon’s contributory negligence, and that the most just solution is a form of what may be called, perhaps somewhat misleadingly, “modified joint liability.” Under *1437such a regime, where the plaintiff is con-tributorily negligent, the fraction of the plaintiffs total damages for which a defendant, who is one of two or more tort-feasors, may be liable to the plaintiff is the same fraction that that defendant’s negligence is of the total negligence of the plaintiff and that particular defendant. In other words, for this purpose only the negligence of the plaintiff and the particular defendant in question are compared.
To avoid double recovery and allow for appropriate contribution, the form of judgment in a case such as this involving a negligent plaintiff and two or more negligent defendants will need to provide for some several liability as well as for some joint and several liability. For example, consider the case of a three-car accident, with each driver, A, B, and C, equally at fault, plaintiff, A, suffering $150,000 total damages, defendants B and C suffering no damages, no party being vicariously liable for the fault of another, and there being no other tort-feasor. In such an instance, modified joint liability would call for A to be awarded a total judgment of $100,000, composed of the following: a $25,000 award in favor of A against B alone; a $25,000 award in favor of A against C alone; and a $50,000 award in favor of A against B and C jointly and severally. To the extent that B were to pay A more than $50,000 under the entire judgment, B would have judgment over for contribution against C; to the extent that C were to pay A more than $50,000 under the entire judgment, C would have judgment over for contribution against B. The precise application of these principles to Simeon’s suit against Lumar and Smith will be discussed in more detail below.
As to the majority’s part VI(b), similar considerations dictate that Mrs. Simeon should not recover ninety percent of her loss of consortium damages from Lumar, but only that fraction — 32/42nds—which Lumar’s negligence is of the combined negligence of Simeon (properly imputed to Mrs. Simeon) and Lumar. The fact that Mrs. Simeon’s negligence claim against Smith ought to have been dismissed on the pleadings (as the majority correctly holds) merely highlights the fact that Smith’s negligence should be irrelevant to Lumar’s appropriate share of Mrs. Simeon’s damages. That should be based simply on a comparison of Lumar’s negligence with that properly imputed to Mrs. Simeon.
A brief overview of some well known tort concepts and their more recent evolution will help place the theory of this partial dissent in context.1
At common law, a fault-free plaintiff could recover all his damages from any one of several concurrent tort-feasors. This was the doctrine of joint and several liability. It was not based on the concept that one of the concurrent tort-feasors should be held liable for the wrong of another. Rather, joint and several liability was premised on the very different concept that each tort-feasor was responsible for all of which his own wrong was a cause. The fault or liability of others — except the plaintiff — was irrelevant. The same reasoning — that a wrongdoer is responsible for all that of which his wrong is a cause— dictated the rule that the plaintiff’s contributory negligence, no matter how slight in comparison to the defendant’s negligence, barred all recovery. This was not because the plaintiff’s negligence made him responsible for the defendant’s wrong, but be*1438cause the plaintiff’s negligence was a cause of all his damages, and he was responsible for all that of which his wrong was a cause. The harshness of this latter result led to the adoption of comparative fault in instances where the plaintiff was contributorily negligent. Under this concept, the complete bar of contributory negligence is lifted, neither party is responsible for all the results of his wrong, each being mutually relieved of a portion thereof, and the loss is apportioned according to the comparative fault of the parties. Where only the plaintiff and one other are at fault, the fault of the injured plaintiff is compared to that of the defendant, and the latter bears that fraction of the loss which his fault is of the total fault of himself and the plaintiff. If the plaintiff and the defendant are equally at fault, the defendant bears half the plaintiffs loss; if the defendant is twice as much at fault as the plaintiff, he bears two-thirds of the plaintiffs loss. If there are additional parties at fault, generally no problem is presented where all are solvent and before the court as defendants. Otherwise, however, the situation may be more difficult, particularly where one of the defendants is insolvent. In such an instance, the basis on which the fault of the plaintiff and that of the defendant or defendants is compared becomes crucial.
Because of the relatively recent adoption of comparative negligence, there are few reported appellate decisions in this area. Generally speaking, only two approaches have been considered. The most widespread approach among the states has been to hold that the solvent defendant is in effect charged with the negligence of the insolvent defendant for purposes of the comparison to be made to the plaintiffs negligence. Referring back to the illustration of the three-car accident, in which each party is equally at fault, this means that if defendant C is insolvent then solvent defendant B must bear two-thirds of plaintiff A’s damages, although A and B are equally at fault. See W. Prosser & P. Keeton, Prosser & Keeton on the Law of Torts § 67 at 475 (1984); F. Harper, F. James, 0. Gray, 4 The Law of Torts § 22.17 at 413-16 (1986); 57 Am.Jur.2d Negligence § 435 at 860 (1971). This appears to me to be an essentially reflexive invocation of common-law joint liability. However, it ignores the fact that common-law joint liability considered the fault of others (except the plaintiff) irrelevant, and did not rest on charging one defendant’s fault to another. It also ignores the fact that comparative fault modified, as between a negligent plaintiff and a negligent defendant, the principle that each party is responsible for all the results of his wrong. Little is said to actually justify comparing the plaintiff’s fault to that of all the defendants in determining the extent of each defendant’s responsibility.2 The result also seems to be influenced by viewing the second approach, pure several liability, as the only alternative. Under pure several liability, each tort-feasor is only liable for the same percentage of the plaintiff’s damages as that tort-feasor’s negligence is of the combined negligence of the plaintiff and all other concurrent tort-feasors. W. Prosser & P. Keeton, supra, § 67 at 475; Sobelsohn, Comparing Fault, 60 Ind.L.J. 413, 455 (1985). Reverting to the hypothetical three-car accident involving equally at fault A, B, and C, the pure comparative fault system results in B’s being responsible for only one-third of A’s damages, even though A and B are equally at fault and A is able to collect nothing from insolvent C. Again, it seems to be assumed that the only alternative to this result is traditional joint liability.
The vice in both conventional joint liability and pure several liability is the same: each looks to the fault of third parties to measure the comparative fault of the plaintiff and the particular defendant, the ex*1439tent of whose responsibility to the plaintiff is under consideration.
A third approach exists, however. For ease of reference, more than descriptive accuracy, it may be called “modified joint liability.” Under it, the relevant comparison for comparative fault purposes is between the plaintiffs negligence and the negligence of the particular defendant, the extent of whose liability to the plaintiff is being fixed. In the mentioned three-car accident example, if C is insolvent, then A is nevertheless able to recover half of his damages from B, since A and B are equally at fault (which is the same result that would have obtained if C had not been at fault). In effect, the risk that C will not compensate plaintiff A (due to C’s insolvency, or C’s immunity from liability, or the fact finder’s determination that what C did was not negligent) is borne by A and B in the respective ratios that the fault of each of them bears to the total fault of both. By contrast, traditional joint liability assigns this entire risk to defendant B, and pure several liability assigns it entirely to plaintiff A.
Though it has become significantly relevant only with the latter-day wide and rapid expansion of comparative fault, the concept of this third solution is itself not of as recent vintage. It was first suggested over fifty years ago by Charles O. Gregory, a professor of law at the University of Chicago. C. Gregory, Legislative Loss Distribution in Negligence Actions, 77-79, 142-48 (1936). Professor Gregory explained the rationale for his proposal:
“At common law joint tortfeasors are virtually guarantors of each other’s solvency so far as concerns the injured plaintiff’s joint judgment for damages; and the introduction of contribution between joint tortfeasors does not affect that situation in the slightest degree. The plaintiff receives his damages at all costs, leaving the defendants to even up the loss between themselves if and as they may and can. But under a comparative negligence statute, where the plaintiff, although negligent, may still recover, the situation is fundamentally different. Here absolutely no reason exists why the defendants, even if they are treated as joint tortfeasors and thus subjected to joint judgment liability for certain purposes, should be made to assume the entire risk of each other’s insolvency with respect to plaintiff’s recoverable damages. For when the plaintiff and the solvent tortfeasor are both negligent, they share the stigma which at common law seems to have furnished the justification for the somewhat arbitrary allocation of this risk on joint judgment debtors. Furthermore, it is quite possible to have a plaintiff who is as negligent as, or more negligent than, either of his defendants, but is still entitled to recover. Under such circumstances, it seems idle to suppose that a joint liability to the plaintiff should carry absolutely the same incidents as the common-law joint judgment; and distribution of the risk of insolvency of one of the joint defendants in accordance with the apportionment of fault would seem to be the only method of administration consistent with the terms of the comparative negligence statute.” Id. at 142 (footnote omitted).
See also Fleming, Foreword: Comparative Negligence at Last — By Judicial Choice, 64 Calif. L.Rev. 239, 252 n. 55, 258 (1976) (praising what is here called modified joint liability and crediting Gregory as its originator).
In a comprehensive study of concurrent fault focusing primarily on Great Britain and Ireland, Professor Glanville Williams advocated modified joint liability as the proper method of apportioning damages when one of the tort-feasors had, for some reason, not been sued. G. Williams, Joint Torts and Contributory Negligence § 110 at 414-20 (1951). Professor Williams implemented this approach when he drafted the Irish contribution statute. See Fleming, 64 Calif. L.Rev. at 252. Other commentators have been supportive. See So-belsohn, 60 Ind. L.J. at 456 (noting that this approach is “[mjore in keeping with the philosophy of proportionate responsibility”); Pearson, Apportionment of Losses Under Comparative Fault Laws — An Analysis of the Alternatives, 40 La.
*1440L.Rev. 343, 354, 362-65 (1980); Note, Reconciling Comparative Negligence, Contribution, and Joint and Several Liability, 34 Wash. & Lee L.Rev. 1159, 1174-76 (1977). But see F. Harper, F. James & 0. Gray, supra § 22.17 at 412-13 (noting “considerable support” for this approach but opposing it because defendants are more able to insure against losses). The Uniform Comparative Fault Act, section 2(d), 12 U.L.A. (West Supp.1987), incorporates this approach, and its principles were judicially adopted by the Missouri Supreme Court in Gustafson v. Benda, 661 S.W.2d 11, 15-16, 21-22 (Mo.1983) (en banc). See also Restatement (Second) of Torts § 886A, comment i (1977) (“If one tort-feasor’s equitable share turns out to be uncollectible it should be spread proportionately among the other parties at fault”). Cf. Duncan v. Cessna Aircraft Co., 665 S.W.2d 414, 429 n. 9 (Tex.1984) (describing this solution as “attractive” but rejecting it, in cases involving strict liability in tort, because of perceived problems of “post-trial jurisdiction and finality of judgments” that would arise in reallocating the share of an insolvent tort-feasor after trial).
We have found one reported appellate decision explicitly applying this approach. Haney Electric Co. v. Hurst, 624 S.W.2d 602 (Tex.Civ.App.-Dallas 1981, writ granted and then dismissed as moot). That case involved a three-car collision. In separate actions, two of the drivers — each later found to be thirty percent negligent — sued the third driver — later found to be forty percent negligent. The cases were consolidated, and one question was whether the defendant (the third driver, found forty percent at fault) should be liable to a particular plaintiff for forty percent of the harm (that is, only the defendant’s share) or seventy percent (that is, the defendant’s share added to the entire thirty percent share of the other plaintiff/tort-feasor, neither of the plaintiffs being also a defendant). The court adopted neither approach and chose instead to hold the defendant third driver liable to each plaintiff for 40/70ths of total damages. That fraction represented the ratio of the defendant’s negligence (forty percent) to the total, of his negligence and the negligence of the party seeking recovery (thirty percent). The court thus placed on defendant a portion of the unsued tortfeasor’s share of fault, but only that portion represented by the ratio of the defendant’s fault (forty percent) to the combined fault of the defendant and plaintiff (seventy percent). The court considered this result mandated not only by the Texas comparative negligence scheme, Tex.Rev.Civ.Stat. art. 2212a (codified as amended at Tex.Civ.Proc. & Rem.Code § 33.001 (Vernon 1986)), but also by “[elementary fairness.” 624 S.W.2d at 612; see also Koonce v. Quaker Safety Products & Manufacturing Co., 798 F.2d 700, 713 n. 17 (5th Cir.1986).
Modified joint liability in this context thus limits the maximum liability to the plaintiff of each individual joint tort-feasor to the amount for which that particular tort-feasor would have been liable to the plaintiff if only the negligence of such party and the negligence of the plaintiff were compared. If the plaintiff is not negligent, this “limit” will necessarily be one hundred percent of all the damages suffered by the plaintiff, and the result will be no different than in traditional joint liability. Accordingly, if modified joint liability is to come into play there must, at the least, be a regime of comparative fault, and the plaintiff and not less than two other actors (neither of whom is vicariously liable for the fault of the other) must each be guilty of causative fault. Assuming, as is the case here, that all these actors are before the court and none has settled with the plaintiff, the operation of modified joint liability may be again illustrated by another three-party example (using a different damages figure and another set of percentages).
If plaintiff A, who suffers total damages of $100,000, is twenty percent negligent, and defendant B is likewise twenty percent negligent, while defendant C’s negligence is sixty percent, then modified joint liability limits B’s maximum exposure to $50,000 (20/40ths of $100,000) and likewise limits C’s maximum exposure to $75,000 (60/80ths of $100,000). However, plaintiff *1441A’s total recovery from both B and C may not exceed A’s total damages ($100,000), less the percentage (twenty percent) thereof which his fault is of the total fault of all parties (A, B, and C) including himself; and thus in this instance, A’s combined recovery from B and C may not exceed $80,000 ($100,000, less $20,000). To effectuate this limitation on A’s overall recovery, as well as the referenced limitations on the maximum exposures of B and C, respectively, the judgment will provide for some several liability and some joint and several liability. Plaintiff A will be awarded a total judgment of $80,000, composed of the following: $5,000 against B alone; plus $30,000 against C alone; plus $45,000 against B and C jointly and severally. Under such a judgment, B’s maximum exposure is $50,-000 ($5,000, plus $45,000), which is half of A’s total damages and corresponds to the fraction which B’s negligence is of the combined negligence of A and B (20/40ths); C’s maximum exposure is $75,000 ($30,000, plus $45,000), which is seventy-five percent of A’s total damages and corresponds to the fraction which C’s negligence is of the combined negligence of A and C (60/80ths); and, A’s maximum recovery is $80,000 ($5,000, plus $30,000, plus $45,000), which is eighty percent of his total damages ($100,000) and reflects the deduction therefrom for his twenty percent share of the combined negligence of all three parties (A, B, and C) at fault.3 Appropriate provision will also need to be made in the judgment for contribution: to the extent B pays A more than $20,000 on the entire judgment, B will have judgment over for the excess against C; and to the extent that C pays A more than $60,000 on the entire judgment, C will have judgment over for the excess against B.4
As applied to the Simeons’ claims against Smith and Lumar, calculation of the appropriate modified joint liability judgment is rendered uniquely complicated by the fact that Simeon’s total damages are $150,000 greater in his claim against Smith than in his claim against Lumar (due entirely to the difference in the pain and suffering awards as respectively made by the jury in the action against Smith and by the court in the action against Lumar). Nevertheless, the task can be accomplished. Assuming Simeon accepts our remittiturs so that his total damages in the claim against Smith are $706,487.50 and his total damages in the claim against Lumar are $556,-487.50, and based on Simeon’s ten percent negligence, Smith’s thirty-eight percent, and Lumar’s thirty-two percent, the liabilities of the parties should be as set out below.
Simeon’s maximum recovery from both Smith and Lumar should be $628,779.93, calculated as follows. Of the $556,487.50 total Simeon damages common to the actions against both Smith and Lumar, Si-meon may recover from all sources ninety percent thereof, or $500,838.75. As to the additional $150,000 of Simeon’s total damages which were fixed only in the claim against Smith, Simeon is entitled to additionally recover from Smith 58/68ths (Smith’s negligence as a fraction of the total negligence of Smith and Simeon) thereof, or $127,941.18. The total of these sums ($500,838.75 plus $127,941.18) is the full $628,779.93 which Simeon is entitled to *1442recover, in part from Lumar alone, in part from Smith alone, and in part from Lumar and Smith jointly and severally. Lumar’s maximum liability to Simeon is for the sum of $423,990.44, which represents 32/42nds (Lumar’s negligence as a fraction of the total negligence of Simeon and Lumar) of the $556,487.50 total Simeon damages common to the Smith and Lumar claims. Smith’s maximum liability is for the sum of $602,592.28, which represents 58/68ths (Smith’s negligence as a fraction of the total negligence of Simeon and Smith) of the $706,487.50 total Simeon damages fixed in the Smith claim. For purposes of contribution, Lumar should have recovery over from Smith if Lumar pays more than $178,-076 (Lumar’s thirty-two percent of the $556,487.50 total Simeon damages fixed in the claim against Lumar), and .Smith should have recovery over from Lumar if Smith pays more than $450,703.93 ($127,941.18, as Simeon’s maximum recovery on the additional $150,000 total damages in the Smith action for which Lumar has no responsibility, plus $322,762.75 constituting Smith’s fifty-eight percent of the $556,487.50 total Simeon damages common to both claims).
To carry the foregoing into effect, the appropriate judgment would award Simeon a total of $628,779.93, composed of: (A) $26,187.65 against Lumar alone; plus (B) $204,789.49 against Smith alone; plus (C) $397,802.79 against Smith and Lumar jointly and severally.5 Lumar, whose maximum potential exposure (if Smith pays nothing to Simeon) is thus $423,990.44 ((A) plus (C)), would have judgment over against Smith for any payments by Lumar in excess of $178,076; Smith, whose maximum potential exposure (if Lumar pays nothing to Si-meon) is thus $602,592.28 ((B) plus (C)), would have judgment over against Lumar for any payments by Smith in excess of $450,703.93.
As to Mrs. Simeon’s loss of consortium claim, the appropriate judgment against Lumar, the only party liable in this respect, is far easier to calculate. Mrs. Simeon’s total damages were $80,000 and, as the majority holds (majority opinion at note 14 and accompanying text) and she does not question, she is properly chargeable with Simeon’s ten percent negligence. The majority holds Lumar liable to Mrs. Simeon for ninety percent of the $80,000 (or $72,-000). But Lumar was only thirty-two percent negligent, and should be liable only for 32/42nds of the $80,000 (or $60,952.38). That is the ratio which the negligence of Lumar bears to the total negligence of Mrs. Simeon (as imputed) and Lumar. Smith’s negligence (fifty-eight percent) is legally irrelevant and Lumar is not vicariously or derivatively liable for Smith’s conduct.
*1443The adoption of comparative fault is generally regarded as an advance in the law, for it avoids the absurd result of an injured plaintiff, who is only slightly negligent, being unable for that reason to recover anything from a defendant whose fault is ten times as great as the plaintiff’s. Should we not likewise avoid the equal absurdity in the opposite direction of allowing a slightly negligent defendant to be liable for eighty-nine percent of the damages of a plaintiff who is ten times more at fault than that defendant.
It must be understood that what is of concern here is not shifting fault or responsibility for damages to or from or among multiple tort-feasor defendants. Indeed, the exact opposite is the case, for the thrust of this partial dissent is that the presence or absence of third-party fault is irrelevant, just as it was at common law. Comparative fault, which comes into play only if the plaintiff is guilty of fault which is a cause of all his damages, says that the plaintiffs fault is to be compared to something. The key question is to what? I suggest that the only logical comparison is to (and only to) the fault of the (or each) defendant from whom the plaintiff seeks recovery (whether the sole defendant or tort-feasor or one of several); if the plaintiff and the defendant are equally at fault, the defendant is responsible for one-half the plaintiffs damages, and the defendant’s responsibility to the plaintiff is neither enhanced nor reduced by whether or not some third person is also guilty of causative fault.
Let us examine a variation of an earlier example. In the three-car accident involving plaintiff A, defendant B, and third driver C, assume that the accident is caused by the speed at which each of the three cars is traveling and that the speed of A and B is each equally negligent and excessive. C (for whose actions B is not vicariously liable) is traveling within the speed limit but at a rate which is arguably too fast due to the slick condition of the road. Why should B be responsible for a larger fraction of A’s total damages if the jury finds that C’s admitted speed was negligent than if the jury does not so find? A finding that C’s speed was negligent does not change the relationship between the negligence of A and B; and it does not imply any enhancement of A’s damages.6 The comparison should only be between A and B. B’s liability to A should be the same (one-half of A’s damages) whether the fault is fifty percent B’s and fifty percent A’s, or forty percent B’s, forty percent A’s, and twenty percent C’s. The majority would have defendant B’s total exposure (for the same damages) go up as his percentage of the total negligence goes down, and even though his negligence relative to that of plaintiff A does not change.7 No one has adequately explained the reason or justice of such a quixotic result.8
Even if, as I firmly believe, modified joint liability is the fairest and most appropriate rule, the question remains whether this Court is free to adopt it or is constrained from doing so by principles of stare decisis and a proper appreciation of the judicial function and our role as an intermediate appellate court. Though these considerations are of the utmost importance, I conclude that they do not preclude our adoption of modified joint liability.
I begin by observing that “[ajdmiralty law is judge-made law to a great extent *1444.." Edmonds v. Compagnie Generale Transatlantique, 443 U.S, 256, 99 S.Ct. 2753, 2756, 61 L.Ed.2d 521 (1979), and that “the Judiciary has traditionally taken the lead in formulating flexible and fair remedies in the law maritime.... ” United States v. Reliable Transfer Co., 421 U.S. 397, 95 S.Ct. 1708, 1715, 44 L.Ed.2d 251 (1975) (replacing divided damages rule with comparative fault). As an admiralty court we are not constricted by some comparative fault statute that might be construed to expressly or impliedly require traditional joint liability. The judicial development of admiralty law is in few situations legislatively strait-jacketed.
For example, though for centuries maritime law required negligence as a basis for recovery (absent unseaworthiness), the federal judiciary in relatively recent years has incorporated strict liability in tort, doing away with the requirement of negligence, in maritime products liability cases. Nor did this development wait upon the initiative of the Supreme Court. As that Court explained in East River Steamship Corp. v. Transamerica Delaval, Inc., 476 U.S. 858, 106 S.Ct. 2295, 2299, 90 L.Ed.2d 865 (1986):
“Absent a relevant statute, the general maritime law, as developed by the judiciary, applies.... Drawn from state and federal sources, the general maritime law is an amalgam of traditional common-law rules, modifications of those rules, and newly created rules.
“We join the Courts of Appeals in recognizing products liability, including strict liability, as part of the general maritime law.” (Emphasis added; footnotes omitted.)
The Supreme Court thus approved what “the Courts of Appeals” had already done, without waiting for word from the high court, to change and adopt a new rule contrary to well established prior practice.
We are aware of no maritime law case where modified joint liability has been considered explicitly and rejected. Indeed, there are relatively few maritime cases where the question might even be raised. What is needed is a case with multiple tort-feasors and a plaintiff who is contribu-torily negligent. Often there may be third-party tort-feasors whose fault is not inquired about because they may be insolvent, or protected by some special immunity or not subject to jurisdiction, and the parties to the suit have simply assumed that the absent party’s fault, if any, was irrelevant to their rights inter se.
In Gele v. Chevron Oil Co., 574 F.2d 243, 251 (5th Cir.1978), our Court simply stated that a contributorily negligent plaintiff could “collect all his damages from one party in the event he is unable to obtain the relative portion of damages from each party at fault.” There was no discussion of alternative methods of apportioning fault. In Empire Seafoods, Inc. v. Anderson, 398 F.2d 204, 217 n. 21 (5th Cir.), cert. denied, 393 U.S. 983, 89 S.Ct. 449, 21 L.Ed.2d 444 (1968), we held that a contributorily negligent plaintiff could recover his damages from a single concurrent tort-feasor defendant. Again, however, no consideration was given to the notion of modified joint liability. In Drake Towing Co. v. Meisner Marine Construction Co., 765 F.2d 1060, 1067 (11th Cir.1985), the Eleventh Circuit held that a contributorily negligent plaintiff could “recover its entire damages, less that proportion attributable to its own fault” from one of two concurrent tort-fea-sors. Drake actually lends some support to the notion of modified joint liability because one of the tort-feasors was not a party and, therefore, it was held error to consider that tort-feasor's comparative share of fault. Instead, Drake held that only the negligence of the plaintiff and the party defendant could be considered. Id. at 1067, 1068 (alleged negligence of un-joined third party “irrelevant,” and liability between plaintiff and defendant must be apportioned “without considering the negligence of” the unjoined third party).9
*1445In Joia v. Jo-Ja Service Corp., 817 F.2d 908 (1st Cir.1987), a maritime case much like this, the First Circuit held a thirty percent negligent defendant jointly liable with a sixty-five percent negligent co-defendant for ninety-five percent of the total damages of the five percent negligent plaintiff. Joia viewed its choices as limited to either making each of the two defendants pay only its own percentage of the total fault of all three parties, or holding the two defendants both jointly liable for ninety-five percent of the plaintiffs total damages. The Court did not discuss modified joint liability. See id. at 915-17. Like us, the First Circuit in Joia found little direct authority on the question. However, it stated that “Edmonds v. Compagnie Generale Transatlantique, 443 U.S. 256, 99 S.Ct. 2753, 61 L.Ed.2d 521 (1979), though[] not controlling, states the admiralty rule applicable here.” 817 F.2d at 916.
In Edmonds, a longshoreman brought a negligence action against the owner of the ship on which he was working. 33 U.S.C. § 905(b). The jury found the plaintiff to be ten percent negligent and the shipowner twenty percent negligent. The plaintiffs employer, against whom the longshoreman’s direct claim was barred by the Long-shore and Harbor Workers’ Compensation Act (LHWCA), 33 U.S.C. § 901 et seq., was found to be seventy percent negligent. The question in Edmonds was whether the shipowner should be liable for only its own share of all fault, twenty percent, or whether it should be liable for ninety percent, its own twenty percent plus the employer’s seventy percent share. The Supreme Court held that the shipowner must pay ninety percent of the plaintiff’s damages. In Edmonds, the Supreme Court
*1446expressed its reluctance to upset the “delicate balance” struck by Congress between the rights of longshoremen, stevedores, and shipowners in the 1972 Amendments to the LHWCA. 99 S.Ct. at 2763. The Court observed that it dealt “with an interface of statutory and judge-made law.” Id. at 2762. The Court noted various inequities that would prevail if the shipowner were liable only for its own twenty percent share of all negligence, id. at 2761, and thought its rule preferable to this alternative. The opinion does not address, much less reject, the possibility of applying the modified joint liability rule here advocated. It. does not even acknowledge the existence of such a third approach.
Analysis of some of the language in Ed-monds may be helpful. In introducing the questions before it, the Supreme Court stated:
“As that [admiralty] law had evolved by 1972, a longshoreman’s award in a suit against a negligent shipowner would be reduced by that portion of the damages assignable to the longshoreman’s own negligence; but, as a matter of maritime tort law, the shipowner would be responsible to the longshoreman in full for the remainder, even if the stevedore’s negligence contributed to the injuries. This latter rule is in accord with the common law, which allows an injured party to sue a tortfeasor for the full amount of damages for an indivisible injury that the tortfeasor’s negligence was a substantial factor in causing, even ¿/the concurrent negligence of others contributed to the incident.” Id., 99 S.Ct. at 2756 (emphasis added; footnotes omitted).
The expression “even if” introducing the concluding clause of each of the above sentences is suggestive of the Court’s assumption that the negligence of the stevedore, or other third party, might normally be thought to reduce the liability of the shipowner (or other primary defendant). The phraseology is certainly at odds with any recognition of the idea that the stevedore’s fault should be put into the equation to increase the percentage of fault for which the shipowner would otherwise be responsible to the negligent plaintiff. All the Court is saying is that the negligence of the stevedore does not decrease the responsibility of the shipowner. That is wholly consistent with modified joint liability.
In a footnote appended to the first sentence in the above passage, the Court remarked:
“We stated the common-law rule in The Atlas [93 U.S. 302, 23 L.Ed. 863 (1876) ] and adopted it as part of admiralty jurisprudence: ‘Nothing is more clear than the right of a plaintiff, having suffered such a loss, to sue in a common-law action all the wrong-doers, or any one of them, at his election; and it is equally clear, that, if he did not contribute to the disaster, he is entitled to judgment in either case for the full amount of his loss.’ ” Edmonds, 99 S.Ct. at 2756 n. 7 (quoting The Atlas, 93 U.S. at 315) (emphasis added).
Obviously, this does not address the question of whether the plaintiff’s negligence is to be compared to that of the (or each) defendant from whom he seeks recovery or to that of all concurrent tort-feasors collectively. ■
In a note called for at the conclusion of the above passage, the Court observed:
“A tortfeasor is not relieved of liability for the entire harm he caused just because another’s negligence was also a factor in effecting the injury. ‘Nor are the damages against him diminished.’ Restatement [ (Second) of Torts], supra, § 879, Comment a.” 99 S.Ct. at 2756 n. 8.
The quoted passage tends to confirm this dissent’s analysis that joint liability is not predicated on one tort-feasor’s being liable for the wrongs of another, but rather on the concept that each tort-feasor is responsible for all the results of his own wrong, and that whether some third party— through fault or otherwise — also contributed thereto is simply irrelevant.10
*1447In a subsequent footnote, Edmonds spoke to the problem of the absence of contribution:
[T]he general rule is that a person whose negligence is a substantial factor in the plaintiffs indivisible injury is entirely liable even if other factors concurred in causing the injury. Normally, the chosen tortfeasor may seek contribution from another concurrent tortfeasor. If both are already before the court — for example, when the plaintiff himself is the concurrent tortfeasor or when the two tortfeasors are suing each other as in a collision case like [U.S. v.] Reliable Transfer [Co., Inc.] [421 U.S. 397, 95 S.Ct. 1708, 44 L.Ed.2d 251 (1975)] — a separate contribution action is unnecessary, and damages are simply allocated accordingly. But the stevedore is not a party and cannot be made a party here, so the Reliable Transfer contribution shortcut is inapplicable. Contribution remedies the unjust enrichment of the concurrent tortfeasor [citation omitted], and while it may sometimes limit the ultimate loss of the tortfeasor chosen by the plaintiff, it does not justify allocating more of the loss to the innocent employee, who was not unjustly enriched. [Citation omitted.] Our prior cases recognize that. Even before Reliable Transfer, we apportioned damages between vessels that collided and sued one another. Reliable Transfer merely changed the apportionment from equal division to division on the basis of relative fault. But we did not upset the rule that the plaintiff may recover from one of the colliding vessels the damage concurrently caused by the negligence of both.” 99 S.Ct. at 2762 n. 30 (final emphasis in original).
Here again, the Court is addressing the question of whether the third party’s fault can diminish the plaintiffs recovery from “the chosen tortfeasor.” Again, the answer is “no.” And “the innocent” plaintiff can recover all his damages from either tort-feasor. But, as before, the question of whose negligence the plaintiffs negligence is compared to is simply not adverted to or even recognized.
There is simply nothing in the language or reasoning of Edmonds which is inconsistent with modified joint liability. Admittedly, the result in Edmonds would have been otherwise under modified joint liability.11 But Edmonds was an LHWCA action, and as we have noted, this was given particular significance by the Court. Moreover, the modified joint liability approach was neither before the court nor even recognized by it. In these circumstances, I would be guided by the principle that “[i]t is not to be thought that a question not raised by counsel or discussed in the opinion of the court has been decided merely because it existed in the record and might have been raised or considered.” United States v. Mitchell, 271 U.S. 9, 46 S.Ct. 418, 419-20, 70 L.Ed. 799 (1926). See also Webster v. Fall, 266 U.S. 507, 45 S.Ct. 148, 149, 69 L.Ed. 411 (1925).12
In sum, I agree with the First Circuit’s Joia that Edmonds is not directly controlling outside of the LHWCA context. But I disagree with any implication from Joia— which itself did not even recognize the existence of the modified joint liability alternative — that Edmonds prevents this Court from adopting modified joint liability in the present case.
*1448Admittedly, modified joint liability is inconsistent with the result or practice in some prior cases. I submit, however, that in a more real world sense there has been no really fixed and largely consistent or widespread conscious understanding or practice on the part of either bench or bar in this respect. Indeed, my impression is that, over the years, most plaintiffs’ lawyers have not (absent some arguable vicarious liability theory) sought to secure findings on third-party fault in the thought that this will increase the percentage of negligence for which the target tort-feasor defendant would otherwise be responsible in the event the plaintiff is found contribu-torily negligent, and that in similar circumstances most damage suit defense counsel have not hestitated, on the basis of any such concern, to attempt to secure findings of third-party fault. I believe that modified joint liability is a reasonable, fair, and just rule which logically flows from and accommodates the interplay of the principles of comparative fault and the responsibility of each party for all the proximate results of his own delicts. The adoption of modified joint liability would represent interstitial legal evolution in the appropriate common-law tradition, a tradition that in the maritime context reaches to the present.
Accordingly, I respectfully dissent from part IVD of the majority opinion and from so much of its part VI(b) as holds Lumar liable for ninety percent, instead of 32/42nds, of Mrs. Simeon’s loss of consortium damages.
. It is recognized that neither Smith nor Lumar has advanced the theory urged in this opinion. Yet I believe that the question — which is purely one of law and is plain on the face of the record — is properly before the panel. Simeon, by cross-appeal (clearly motivated by fears of Lumar’s insolvency), challenges the district court's limitation of his recovery against Smith to fifty-eight percent of his total damages. We properly find this limitation is in error and we must accordingly tell the district court what the appropriate percentage is. Surely we will not tell the district court to apply what we believe to be a legally erroneous percentage. Similarly with respect to the loss of consortium claim, Lumar appeals, contending that assessing it ninety percent of the damages was error, in light of the negligence percentages found. If we conclude that Lumar is correct — that, as it claims, ninety percent is too much — then it makes little sense to affirm simply because Lu-mar has sought too great a reduction — to thirty-two percent rather than to 32/42nds.
. I do not address the appropriateness of such a comparison where its purpose is to determine whether, under a modified comparative fault regime, the plaintiffs contributory negligence is sufficiently great to bar all recovery, or all recovery from a particular one of several defendants. See, e.g., Mountain Mobile Mix, Inc. v. Gifford, 660 P.2d 883, 886-87 (Col.1983) (reviewing authorities on this point).
. B’s sole liability for five percent ($5,000) of A’s total damages is arrived at by subtracting C’s seventy-five percent ($75,000) maximum exposure percentage from A’s eighty percent ($80,-000) maximum recovery percentage. Similarly, C’s sole liability for thirty percent ($30,000) of A’s total damages is calculated by subtracting B’s fifty percent ($50,000) maximum exposure percentage from A’s eighty percent ($80,000) maximum overall recovery percentage. The forty-five percent ($45,000) of A’s total damages for which B and C are jointly and severally liable is calculated by subtracting the sum of B's five percent ($5,000) sole liability and C's thirty percent ($30,000) sole liability — which aggregate thirty-five percent ($35,000) — from A’s eighty percent ($80,000) maximum overall recovery (eighty minus thirty-five equals forty-five).
. B’s entitlement to contribution from C for ■ payments by B in excess of twenty percent ($20,000) of A’s total damages corresponds to B’s percentage (twenty percent) of the combined fault of A, B, and C. Similarly, C’s entitlement to contribution from B for payments by C in excess of sixty percent ($60,000) of A’s total damages corresponds to C’s percentage (sixty percent) of the combined fault of A, B, and C. This is no different than contribution under conventional joint liability.
. The (A), (B), and (C) amounts are calculated as follows:
(A) The $26,187.65, for which Lumar is solely responsible, constitutes the difference between (i) the $500,838.75', which represents the ninety percent maximum that Simeon is entitled to recover from all sources out of the $556,487.50 total damages common to both the Smith and Lumar claims, and (ii) the $474,651.10, representing the maximum portion (58/68ths) of that $556,487.50 for which Smith could be responsible to Simeon ($500,838.75 minus $474,651.10 equals $26,187.65).
(B) The $204,789.49, for which Smith is solely responsible, is calculated in three steps. First, with respect to the $150,000 total Simeon damages which are entirely inapplicable to the claim against Lumar, Smith (and Smith alone) is responsible to Simeon for $127,941.18, which is 58/68ths thereof. Second, the amount of $76,-848.31 is calculated with reference to the $556,-487.50 total damages common to the Smith and Lumar claims; Simeon is entitled to recover ninety percent of this $566,487.50, or $500,-838.75, from all sources, and up to $423,990.44 thereof (32/42nds) from Lumar. The $76,-848.31 represents the difference between the $423,990.44 (the maximum recoverable from Lumar) and the $500,838.75 (the maximum recoverable from all sources respecting the $556,-487.50). Third, the $127,941.18 (step one) and the $76,848.31 (step two) are added together to give the total of $204,789.49 awarded against Smith alone.
(C)As to Lumar, this $397,802.79 for which it has joint liability with Smith, represents the difference between Lumar's maximum potential liability to Simeon of $423,990.44 (32/42nds of the total $556,487.50 Simeon damages in the Lumar claim) and the $26,187.65 ((A) above) for which Lumar alone is liable. As to Smith, this $397,802.79 for which it has joint liability with Lumar, represents the difference between Smith's maximum potential liability to Simeon of $602,592.28 (58/68ths of the $706,487.50 total Simeon damages in the Smith claim) and the $204,789.49 ((B) above) for which Smith alone is liable.
.The same point can be made if the illustration is changed so that, for example, C’s conduct contributing to the accident is his admitted failure to see the clearly visible cars driven by A and B; the evidence is conflicting as to whether C’s failure in this respect stems from his negligent inattention or, as he claims, from his then suffering a sudden, momentary and unforeseeable fainting spell. A finding that C was negligently inattentive does not change the relationship between the negligence of A and B or imply any enhancement of A’s damages.
. Moreover, if C is found at fault, A has an additional source for potential recovery.
. Pure several fault is also unjust in this scenario. If B can take any advantage of C’s being at fault, it should only be as against C (by contribution), not as against A (by lowering B's exposure to A), for the addition of C’s fault does not change the ratio of fault between A and B, and B’s fault remains a cause of all A’s damages whether or not C also is guilty of causative fault.
. In Drake, the court indicated that a different result (where the third-party settlement payment would reduce the plaintiff’s recovery by the proportion which the third party’s fault bore to the total fault of the plaintiff, the defendant, and the third party, rather than on a dollar-for-*1445dollar basis) might be reached if the defendant had impleaded the third party (in which event, the extent of the third party’s negligence would presumably be relevant). Id., 765 F.2d at 1068. However, in Self v. Great Lakes Dredge & Dock Co., 832 F.2d 1540, 1544-48 (11th Cir.1987), the Eleventh Circuit applied the Drake approach (dollar-for-dollar credit of the third-party settlement payment), even though the third party was before the court on the defendant's cross-action and its percentage of fault was ascertained. Self regarded this holding as mandated by Ed-monds, though it recognized it to be contrary to our holding in Leger v. Drilling Well Control, Inc., 592 F.2d 1246 (5th Cir.1979), and in post-Edmonds cases where we followed Leger. See Self, 832 F.2d at 1548 & n. 6. We have indeed followed Leger (settlement reduces recovery by proportion of settling party fault) in post-jEd-monds cases, both maritime, e.g., Martin v. Walk, Haydel & Associates, Inc., 742 F.2d 246, 249 (5th Cir.1984); Bass v. Phoenix Seadrill/78, Ltd., 749 F.2d 1154, 1157, 1159 (5th Cir.1985); see also In re Incident Aboard D/B Ocean King, 813 F.2d 679, 689 (5th Cir.1987), and otherwise, e.g., Diggs v. Hood, 772 F.2d 190, 196 (5th Cir.1985) (relying on Leger). On the other hand, in our recent decision in Hernandez v. M/V Rajaan, 841 F.2d 582, 591 (5th Cir.1988), we applied Self at the behest of the defendant, who successfully sought a dollar-for-dollar credit for the settlement amounts paid the plaintiff by the third-party defendants, even though there was apparently no finding of the percentage of fault attributable to the third-party defendants.
It is not necessary for this dissent to take a position as to whether the Leger rule or the Self rule represents binding precedent or is preferable. All of these cases concern the special problems attendant on how to credit the settlement of one of several tort-feasors. That is not involved here.
Under the Leger approach, if plaintiff A settles before trial with concurrent tort-feasor C, and at trial A is found to have $150,000 total damages and to be one-third negligent, while C and defendant B are also found to each be one-third negligent, then B is held liable to A for $50,000, as the settlement, whatever its amount, is treated as discharging one-third of A’s total damages. In that context, modified joint liability has no role to play. It is appropriate to regard A’s settlement with C (regardless of its amount) as encompassing both the sole liability which C would otherwise have had for $25,000 and C’s half of the joint liability (with B) for $50,000. A and C have voluntarily agreed on this when either could have insisted on trial, and so there is no reason either should be able to question it (A to seek more on account of C’s fault, C to seek contribution); B may not justly complain that C paid the wrong amount, for it was C, not B, that did the paying, and B is treated no worse than if there had been no settlement.
Under the Self approach, it is presumably not proper to find whether or not the settling party is negligent, or, if so, what its percentage of negligence is. This suggests that the relevant comparison at trial is between the negligence of the plaintiff and the nonsettling defendant only. That was apparently the case in Hernandez, 841 F.2d at 591, and is fully consistent with modified joint liability. In Self, the plaintiff was not negligent, so modified joint liability simply could not come into play.
None of these cases, however, in any way addresses the concept of modified joint liability.
. As the Edmonds footnote subsequently points out, this assumes that the injury in question is not divisible. Edmonds, 99 S.Ct. at 2756 n. 8. I *1447make the same assumption throughout this dissent.
. Under modified joint liability, the Edmonds shipowner would have been liable for two-thirds (20/30ths), rather than ninety percent, of the injured plaintiff longshoreman’s total damages, as there the negligence percentages were: twenty percent, shipowner; ten percent, plaintiff; seventy percent, stevedore.
. In Webster, the Court stated: "We do not stop to inquire whether all or any of them [prior Supreme Court decisions cited by appellant] can be differentiated from the case now under consideration, since in none of them was the point here at issue suggested or decided. The most that can be said is that the point was in the cases if any one had seen fit to raise it. Questions which merely lurk in the record, neither brought to the attention of the court nor ruled upon, are not to be considered as having been so decided as to constitute precedents.” Id., 45 S.Ct. at 149.