The sole issue presented on plaintiffs’ appeal as permitted by leave of this court pursuant to Rule 105, Rules of Civil Appellate Procedure, is whether in this negligence action the trial court’s order granting a new trial for insufficiency of evidence may properly be limited to a retrial of the single issue of the percentage of causal negligence attributable to each tortfeasor.
The issue arises out of a jury trial of an action for personal-injury damages resulting from a collision between two automobiles in an uncontrolled intersection in the city of St. Paul. The issues of negligence, causation, and the comparison of causal negligence attributable to each driver (plaintiff Joseph Raymond Juvland and defendant Peter Donald Mattson) were submitted in the form of a special verdict as authorized by our new comparative negligence statute. Minn. St. 604.01, subd. I.1 The jury awarded the injured plaintiff driver $11,000 damages and found *367both drivers causally negligent, attributing 90 percent to defendant driver and 10 percent to plaintiff driver. Thereafter, defendants moved the trial court for an order amending the apportionment of causal negligence between the drivers to 50 percent each, or in the alternative for a new trial limited solely to the “issue of division of negligence” between the drivers. The trial court granted the latter alternative of defendants’ motion upon the ground that the jury’s apportionment of negligence was “not justified by the evidence and is contrary to the overwhelming weight of the evidence.” Because the question is one of first impression which we deem should be resolved in the interest of the administration of justice, we granted plaintiffs’ request for discretionary review.2
We hold that the trial court erred in too narrowly limiting the issue to be retried and remand for a retrial of the general issue of liability.
The separate questions submitted in this case pertaining to the negligence of each driver were identically worded:
“At and immediately prior to the accident was [named driver] negligent in the operation of his automobile.”
The question called for a determination of the ultimate fact issue of negligence and did not, as is sometimes the practice, require answers to separate questions concerning each particular respect in which the evidence would support a finding of negligence, e. g., in the following form:
At and just prior to the collision, was [the driver of the vehicle] negligent with respect to:
(a) Lookout? Answer: ............
*368(b) Control? Answer: ............
(c) Speed? Answer: .............3
A similar question is then submitted concerning causation, followed by a comparative negligence question in the case of multiple tortfeasors. In Wisconsin, which has long been the recognized leader in applying the law of comparative negligence, this was the only authorized method of submission prior to 1961. Thereafter, the Wisconsin Supreme Court pursuant to its rule-making power amended a statute on the subject to give the trial court discretion to submit an ultimate fact question of the type employed in this case. Baierl v. Hinshaw, 32 Wis. (2d) 593, 146 N. W. (2d) 433.
In this case, defendant driver was approaching the intersection from the south, traveling in a northerly direction, and the plaintiff driver was approaching from defendant’s right, driving in a westerly direction. Plaintiff’s speed was 20 miles per hour and defendant’s, 20 to 25. It was daytime, the day was clear and dry, and the collision occurred within the intersection. A perusal of the record indicates that there was evidentiary support for a permissible finding of causal negligence on the part of defendant driver with respect to lookout, failure to yield the right-of-way, failure to reduce speed, and possibly control; and as to plaintiff driver, with respect to lookout, control, and possibly speed.
Determination of the percentage of negligence attributable to each driver presupposes a determination of the particular respect in which each driver was causally negligent. The special verdict returned by the jury in this case in no way discloses that determination. Thus, on retrial, if the ultimate fact issue of causal negligence of each driver were to be given binding effect, the trial court would be limited in merely declaring that the causal negligence of each driver had been previously established without specifying in what particular respect each was causally *369negligent. It is thus clear that the issue of the percentage of negligence attributable to each driver is not separate and distinct from the issues of the particular negligence claimed and its causal effect. In our opinion, therefore, the order limiting the retrial to this narrow issue is neither feasible nor practical, in the sense of saving litigation time and expense. The question is not dissimilar to the one presented in Koenigs v. Werner, 263 Minn. 80, 85, 116 N. W. (2d) 73, 76, where the trial court ordered a new trial limited to the issue of proximate cause. Holding the order nonappealable of right, we observed:
“* * * [T]he practice of permitting a partial new trial should not be followed unless it clearly appears that the issue to be retried is distinct and separable and that the trial of that issue alone may be had without injustice.”
We are fully mindful that Rule 59.01, Rules of Civil Procedure, provides that a new trial may be granted “on all or part of the issues.” Although this broad discretionary authority is intended to prevent retrials of any issue properly and fairly decided, it is obvious that such authority could not possibly be used to further limit a retrial to the question of the percentage of negligence attributable to only one of multiple tortfeasors following a verdict where the combined negligence did not equal 100 percent, even though the jury assessment against the other tortfeasors was supported by the evidence. While it may be conceivable that a trial court could limit the retrial of the issue of apportionment of the negligence in a case where the special questions submitted in the former trial determined each respect in which the tortfeasors were alleged to be negligent, a limited retrial without the risk of prejudice to one or the other of the parties is not feasible where the question submitted related to the ultimate fact of causal negligence.4 For example, on retrial of the limited *370issue in this case, plaintiff driver, merely from the court’s instruction that his negligence had been established, would be exposed to a prejudicial inference that the percentage of negligence attributable to him was greater than the 10 percent found by the first jury simply because of the proper and natural tendency to treat the tortfeasors as equally at fault both at the outset of trial and jury deliberations. This risk of prejudice would be especially real where, as here, plaintiffs, not without justification in our opinion, argue that the evidence would have supported a finding that plaintiff driver was in no respect negligent.
Defendants rely on several Wisconsin cases which, to be sure, include language that would lead one to believe that it is a practice in that state to limit retrials to the narrow issue of the apportionment of negligence only. Firkus v. Rombalski, 25 Wis. (2d) 352, 130 N. W. (2d) 835; Caldwell v. Piggly-Wiggly Madison Co. 32 Wis. (2d) 447, 145 N. W. (2d) 745; Lund v. Thorne, 266 Wis. 239, 63 N. W. (2d) 317. In these cases, the Wisconsin court acknowledges its reluctance to upset the apportionment of negligence except in unusual cases where it appears as a matter of law that the plaintiff’s negligence equaled or exceeded that of defendant, or where the allocation of negligence between the parties is so disproportionate when viewed in the light of the acts of negligence charged as to compel the conclusion that justice has miscarried. None of these cases, however, appear to be a review of a trial court’s order limiting the issue to be tried, as in this case. Moreover, a careful reading of these cases indicates that while a retrial of the narrow issue of apportionment appears to be ordered, the appellate court expressly directs that evidence relating to the issue of general liability may be introduced for the purpose of determining apportionment. As we understand, the trial bench of Wisconsin regards such a remand as requiring a retrial of the general issue of liability because, as a practical matter, the jury must first determine the particular respect in which each party is causally negligent before a comparison of *371fault can feasibly be translated into a percentage of causal negligence.
Plaintiffs urge that we review the merits of the trial court’s order, which now will require a retrial of the general issue of liability. We decline to do so because that scope of review was not contemplated by our order granting discretionary review and could not fairly have been anticipated by defendants. This is not to say that plaintiffs would not be justified before the trial in moving for a reconsideration by the trial court of the order as now amended.
Reversed and remanded.