OPINION OF THE COURT
Arthur W. Lonschein, J.In this medical malpractice action, the jury reported a plaintiff’s verdict against both defendants, in the total sum of $2,930,000, apportioning liability on the basis of 75% against the defendant Levin and 25% against the defendant Booth Memorial Medical Center. When the jury was polled, it was found that the verdict was not unanimous on any of the issues. Juror number 4 voted for liability on the hospital, but did not join in the finding of liability as to Dr. Levin, in the measure of damages, or in the apportionment of liability. Juror number 5 did not join in the finding of *465liability as to Booth Memorial, but did join in the other findings, including the apportionment of damages. The jury had been instructed that before it could report its verdict, at least five jurors had to agree, and that it had to be the same five jurors in agreement as to all of the issues. After the verdict was returned, the court was not requested to send the jury back for further deliberations.
Both defendants now request the court to set aside the verdict, on a variety of grounds. These can be summarized as follows:
(1) The jury clearly disregarded the court’s instructions in reporting its verdict without the agreement of the same five jurors on all issues, and this violation of instructions, in and of itself, warrants a new trial.
(2) The verdict is insufficient since less than five jurors agreed on all issues, as well as internally inconsistent due to the votes of juror number 5.
(3) The court’s charge was unfair in that the evidence was marshaled in an improper manner and allowed the jury to draw certain unwarranted inferences.
(4) This verdict was excessive and was influenced by certain testimony which should have been stricken.
THE FORMAL SUFFICIENCY OF THE VERDICT
The threshold question which must be addressed is whether the degree of agreement actually reached and reported by this jury is formally sufficient to constitute a verdict which may be accepted by the court. Left for later discussion are the possible grounds for setting the verdict aside.
The court believes that as to all issues except the apportionment of liability between the defendants, a formally sufficient verdict has indeed been returned.
CPLR 4113 (subd [a]) provides that the concurrence of five sixths of the jurors is enough for the jury to render its verdict. The statute is silent on the question of whether, in a case which involves more than one issue or more than one defendant, the same group of jurors must agree on all of the issues. The court had instructed the jury that indeed it must be the same five jurors in agreement before a *466verdict could be rendered. As will be shown below, this procedure is necessary in any case which involves a possible apportionment of liability between defendants, in order to preclude the possibility of a verdict depending on one or more jurors voting inconsistently.
The reported decisions in this area are in conflict. In Murphy v Sherman Transfer Co. (62 Misc 2d 960) a jury of 12 returned a plaintiff’s verdict as follows: a core group of nine jurors agreed on liability and damages. They were joined by one of the remaining jurors on liability (making 10 on that issue) and by a different remaining juror on damages (making 10 on that issue). The Appellate Term reversed, holding (p 961) that “A general verdict is an indivisible entity and it cannot readily be separated into its component elements”.
This case appears to have been overruled by the Appellate Division, First Department, in Bichler v Lilly & Co. (79 AD2d 317). That case involved a unanimous general verdict by a jury of six. The general verdict was accompanied, however, by written interrogatories. There was a core group of only four jurors who agreed on all of these. One of the other jurors agreed with the core group on the first interrogatory (making five) and a different juror joined the core group on the second interrogatory (again making five). The Appellate Division, without mentioning Murphy (supra), ruled that there was no constitutional defect in the verdict and upheld it. The votes of both dissenting jurors on the interrogatories were inconsistent with their votes on the general verdict, but this issue does not appear to have been preserved for review, and in any event the court did not discuss its ramifications.
In Reed v Cook (103 NYS2d 539) there were three defendants. The jury of 12 found unanimously against one defendant, by 10 to 2 against a second, and by a different 10 to 2 in favor of the third. This was before Dole v Dow Chem. Co. (30 NY2d 143) and there was no claim for apportionment of liability. The court held that since the case consisted essentially of three separate cases (plaintiff against each of the three defendants) which could readily have been tried separately, there was no need for the same *46710 jurors to agree on all of the issues, and accepted the verdict.
In Forde v Ames (93 Misc 2d 723) the court was faced with essentially the same facts as in Murphy v Sherman Transfer Co. (supra) and accepted the verdict, reasoning that such a pattern of votes would have been proper in a bifurcated trial. The court stated (p 725) that “[t]he appropriate assumption in a case such as this is that when a juror is outvoted on the question of liability he will accept the outcome and continue to deliberate with the other jurors honestly and conscientiously to decide the remaining issues”.
Finally, in Orens v Secofsky (60 AD2d 866) the Appellate Division was faced with a liability verdict in a bifurcated trial where different combinations of five jurors had returned verdicts against one of the two defendants and in favor of the other. The court (p 867) termed the failure of five jurors to agree on all issues a “disquieting factor”. It also seems that the verdict consisted of answers to two interrogatories as to each defendant, and that the jury’s response to the interrogatories as to the exonerated defendant were inconsistent. The court (p 867) did not rule on the propriety of the verdict on any grounds, however, since the parties had stipulated to the procedure which gave rise to the problem, and had, therefore, “charted their own course”.
These cases make it clear that a verdict returned with the concurrence of a core group of less than five jurors, supplemented by different remaining jurors on different issues, is not for that reason alone defective on constitutional grounds. None of them, however, squarely confronted the issue now before the court, whether such a voting pattern was proper in a case in which the “swing” jurors were voting inconsistently on different issues. The issues involved in Murphy, Forde and Reed (supra) were all logically independent, and the inconsistency issue was not preserved for review in either Bichler or Orens (supra).
The court noted above that to allow this pattern of voting in a case involving an apportionment is to invite the problem of a verdict depending on a juror’s internally *468inconsistent votes. This arises as follows: a juror cannot consistently dissent from a finding of liability as to any defendant and, at the same time agree to an apportionment of liability which places some of the blame on that defendant. If only one juror dissents there is no problem, since his votes can be ignored and the verdict returned by the remaining five. If, however, there are two dissenters, in order for there to be a total of five votes on each issue all that is necessary is for them to disagree on the liability of different defendants, while one of them joins in the apportionment. Clearly, the second dissenters’ votes are inconsistent.
Analysis of the possible inconsistencies introduced by allowing different jurors to dissent on different issues rapidly becomes unmanageable as the number and character of the issues increase. Consider, for example, a case where there are two defendants, one of whom is sued on two distinct theories, and where there must be an apportionment of liability. Here there are four liability issues, and there can be as many as four dissenters, one on each issue, with the concurring jurors adding up to five on each issue. If there are four (or even three) dissenters, at least one of them must be voting inconsistently.1 The possibilities for inconsistency increase geometrically if one considers the cases where individual jurors dissent on more than one issue. 1
Always, however, the difficulty is introduced by the presence of an apportionment issue. An apportionment requires a refinement of the jurors’ earlier votes on the liability of the defendant. It fixes, not merely the existence or lack of liability, but the degree of liability.
It is of course desirable that the jury’s possible votes be structured in such a way as to avoid internally inconsistent voting, and this cannot be done if a juror is allowed to vote “No Liability” as to a given defendant and still participate in an apportionment which contradicts that vote. For that reason, the court feels strongly that it was correct when it instructed the jury that if only five of them agreed on the verdict, it had to be the same five on all issues.2
*469Upon close examination of the votes of the individual jurors as revealed by the polling of the jury it is clear that such defect as exists in this verdict must inhere in the inconsistent votes of juror number 5. That juror’s votes were inconsistent in that he did not agree with the jury’s finding that there was “negligent treatment by Booth Memorial Medical Center which was a proximate cause of David Cohen’s injury” (that is, he did not find any degree of liability as to the hospital) and that he then turned around and concurred in the apportionment, placing 25% of the total blame on the hospital. These votes cannot be reconciled, and the court cannot choose between them and sustain one of them. In the court’s view, they must both be eliminated from consideration. This juror’s votes on the other issues are not inconsistent and can remain.
Considering the jury’s votes after elimination of juror number 5’s inconsistent votes, there are five jurors in agreement that Dr. Levin is liable (juror numbers 1, 2, 3, 5 and 6), that Booth Memorial is liable (numbers 1, 2, 3, 4 and 6) and that the total damages should be fixed at $2,930,000 (numbers 1, 2, 3, 5 and 6). As to the apportionment, the result is different. Here there is no fifth vote to add to that of the core group of four (numbers 1, 2, 3 and 6). Juror number 4 dissented, and juror number 5’s vote cannot be considered as discussed above.
Must the verdict be discarded in toto merely because there is no agreement on the apportionment? The court thinks not. The jury found, with a sufficient degree of unanimity, that both defendants were separately liable to the plaintiff in a specific amount. As between the defendants and the plaintiff, the jury’s verdict was formally sufficient. If the plaintiff had sued the defendants separately, there would be no quarrel with this result. (See Reed v Cook, 103 NYS2d 539, supra.)
*470The apportionment is strictly a matter between the defendants. The plaintiff has no interest in it. It is as to the apportionment, and the apportionment alone, that the jury failed to reach a sufficient degree of unanimity to return a verdict. The court can see no reason why the plaintiff’s case against the defendants should be retried merely to resolve this issue. Indeed, it has been held that a trial between two joint tort-feasors to fix apportionment of liability may be had separate from the trial in which their liability to the injured party was adjudicated. (Dole v Dow Chem. Co., supra; Westchester Light. Co. v Westchester County Small Estates Corp., 278 NY 175.) Consequently, the cross claims are severed, a mistrial is declared as to them, and a new trial ordered. On the retrial of this issue, if either defendant cares to pursue the matter, the sole task of the trier of the facts will be to assign percentages, since the liability of both defendants has been established.
. The proof of this is left as an exercise for those who care to pursue the matter.
. The court also feels that the same ruling should apply to cases without an apportionment, on the theory that it is more difficult to convince a group of five to agree on all *469issues, and the difficulty of a plaintiff’s gaining a verdict should not differ from case to case depending on whether the defendants are entitled to, or have asked for, an apportionment. Also, and despite the ruling in Bichler v Lilly & Co. (supra) it is still possible to make an argument for the applicability of the result reached here to all cases, on the general grounds that CPLR 4113 (subd a), which allows five-sixths verdicts, while it is authorized by the State Constitution (art I, § 2) is still in derogation of the common-law requirement of unanimity and hence must be construed strictly. Be that as it may, the court’s opinion today rests entirely on the pragmatic desire to avoid precisely the kind of problem which has arisen here.