The opinion of the court was delivered by
Conford, P. J. A. D.In this condemnation case, tried to a jury, involving two separate properties of the same owner, bnt differently located, the trial judge vacated the jury verdict as to both properties on the ground that the testimony of the State appraisal expert, Walter H. Geary, “was wilfully false, * * * was of a substantial natnre, conld have affected the creditability of the witnesses” of the landowner, “impeded cross-examination * * * as to a comparable property * * * and could have had an * * * influence” upon the jury’s verdict. We granted leave to appeal.
Essentially, the trial court determined that the named witness testified willfully falsely in that he failed on direct examination to cite or nse as a “comparable sale” a sale from Mapland Corporation to Jersey Central Power and Light Company of nearby property, and that he explained on cross and re-direct examination as his reason for failure to use the sale that the buyer was under stress of a special need to obtain land in the particular area for use as a substation to replace an existing sub-station being eliminated by a new jughandle on the highway. The basis for the conclusion of willful falsity of the foregoing was the circumstance that the same .witness, in testifying for the State a few days later in another condemnation case but before a dif*549ferent judge, as to the value of property of one Kaupelis located in the general area of the Probaseo and Mapland properties, had there cited and relied on the Map]and-Jersey Central sale in arriving at his valuation.
On the argument of the motion to vacate the judgment before the trial court the State requested the opportunity to have Geary explain by deposition his use of the Mapland sale in the Kaupelis ease in the light of his failure to use it in the instant case, no inquiry as to that question having been addressed to him in either trial, but the request was denied. We ordered such a deposition to be taken, in the exercise of our original fact-finding jurisdiction, and will refer to it later, but we are of the view that the action of the trial court was unjustified without regard to the contents of the deposition.
We emphasize at the outset that the jury verdict was not set aside on the ground that it was against the weight of the evidence, notwithstanding that was one of the grounds of the motion to set aside the verdict, but only for the assertedly willfully false testimony of the State’s appraisal expert.
It is clear that willfully false testimony of a witness may in some circumstances justify setting aside a verdict or judgment. But there are stringent qualifications on the appropriate exercise of the power to vacate on that ground. They have been stated as follows in Shammas v. Shammas, 9 N. J. 321, 330 (1952) :
“. . . Perjured testimony that warrants disturbance of a final judgment must be shown by olear, convincing and satisfactory evidence to have been, not false merely, but to have been willfully and purposely falsely given, and to have been material to the issue tried and not merely cumulative but probably to have controlled the result. Further, a party seeking to be relieved from the judgment must show that the fact of the falsity of the testimony could not have been discovered by reasonable diligence in time to offset it at the trial or that for other good reason the failure to use diligence is in all the circumstances not a bar to relief. Balip Automotive Repairs, Inc. v. Atlantic Casualty Ins. Co., supra. Clearly the necessity to satisfy these tests before the judgment may be disturbed is itself a deterrent to repeated litigation of the same factual issues.” (Emphasis ours).
*550See Balip Automotive Repairs, Inc. v. Schroeder, 8 N. J. Super. 238 (App. Div. 1950), aff’d, Balip Automotive Repairs, Inc. v. Atlantic Casualty Insurance Co., 7 N. J. 152 (1951); Cermak v. Hertz Corp., 53 N. J. Super. 455 (App. Div. 1958), aff’d o. b. 28 N. J. 568 (1959).
In the present ease, the allegedly false testimony being that of an expert witness, the general rule aforestated must be further refined to reflect the concept that to be accounted perjurious the rendition by a witness of an opinion must be shown to have been adduced in bad faith, i- e., that the opinion was one which in fact the witness did not entertain. State v. Sullivan, 25 N. J. Super. 484, 490 (App. Div. 1953), certif. den. 13 N. J. 289 (1953); State v. Sullivan, 24 N. J. 18, 43 (1957) (per Weintraub, J., dissenting).
Here, willful falsity by Geary could be predicated only on either (1) falsity of the witness’s testimony in this case, on cross and re-direct examination, that in effect the Mapland sale was made under the influence of some compulsion on the part of the buyer or (2) that Geary’s omission of use of the sale when testifying on direct examination was in bad faith, his true, concealed opinion being that the sale did have a bearing upon his evaluation of the parcels under appeal.
As to the first hypothesis of falsity, the theory of compulsion in the sale could be argued both ways, and Geary was vigorously cross-examined on the point. He said discussions with Jersey Central officials supported his view of the sale. If the property owner disputed the factor of compulsion she could have offered rebuttal on the point by her own expert witness, who did cite the sale as a comparable in his testimony (but only as to the “small” Probasco parcel, not the “large” one; see infra). Thus (1) subject to the effect to be given the transcript of Geary’s testimony in the Kaupelis ease, as to which see infra, the trial court had no fair basis for finding Geary’s testimony willfully false as to the fact of an element of compulsion in Jersey Central’s purchase; and (2) there was no satisfaction of the requirement for relief that the de*551fendant should have been without the means to refute the fact at the trial. Shammas v. Shammas, supra (9 N. J. at 330).
In dealing with the second potential theory of falsity of the testimony —• bad faith in failure as an expert witness to use the Mapland sale in evaluating the property under condemnation — some further explanation of the facts is necessary. Gearv’s failure below to use the Mapland sale is pertinent in the present context only to his valuation of the small Probasco piece, which is situated on the east side of Millhurst Tennent Eoad, a short distance south of Eoute 33. He was challenged by defendant on the sale only as to the small parcel. None of the witnesses on either side (the property owner had three experts) cited the sale as pertinent to the valuation of the large Probasco parcel which occupies the northwest corner of Eoute 33 and Millhurst Tennent Eoad, apparently because of the great size and depth of the latter. The Mapland parcel was on the east side of Millhurst Tennent Eoad, but became connected with Eoute 33 by a 35-foot easement grant which Jersey Central negotiated at about the time it purchased from Mapland. The Kaupelis property is on the south side of Eoute 33, some distance east of Millhurst Tennent Eoad, and directly across the highway from the Jersey Central easement. Eoute 33 was zoned commercial on both sides, but the small Probasco parcel was zoned residential at the time of the taking, being a vacant lot between two old residences. Geary testified that its highest and best use was for residential purposes. The liaupelis property was the site of a tavern when taken.
As a matter of law and good real estate appraisal practice, the comparability of a sale to a subject property is a matter of the judgment of the individual appraiser, and different appraisers can and frequently do have good-faith differences of opinion. The comparability of a sale to a given subject property is not necessarily a mandate of its comparability to another subject property, even if located in the same general area. Differences in the highest and best use, the actual or *552intended use, the zoning of the respective subject parcels, or other factors may legitimately lead an expert to the conclusion that the sale has probative valuation weight in relation to the one subject parcel but not to the other.
The essential criterion in an expert’s decision to use a sale as a comparable is a substantial similarity of conditions. See the discussion in Moorestown Tp. v. Slack, 85 N. J. Super. 109, 113-114 (App. Div. 1964), certif. den. 43 N. J. 452 (1964). But the expert may well discount dissimilarity of one or more conditions, if there is siuch similarity as to others so that, after adjusting for the dissimilarities, there remains such educational value in the sale price as to help the appraiser in valuing the subject property. Id., at 114. The differences requiring consideration, and, perhaps, adjustment, may include those as to physical aspects, degree of improvement, actual or highest and best use, location, time or date of sale and the degree of deviation from the willing buyer-willing seller concept (e. g., as here, pressure on the buyer to buy or on the seller to sell). Id., at 115.
The gist of the reason for the trial court’s decision that Geary here gave willfully false testimony was that he used the Mapland sale as a comparable in the Kaupelis ease but not in this one. But he was not questioned in Kaupelis concerning any factor going to the concept of willing buyer-willing seller as to that sale. Eor reasons hereinabove explained, Geary may well have considered that the Mapland sale was usable in valuing the Kaupelis property, after adjusting for the factor of compulsion in the sale, because of the outweighing factors of similarity of character of special use (commercial or industrial) and direct access to Route 33, with its high commercial location-value factor, and nevertheless rejected the sale in valuing the small Probasco piece because of tire residential zoning and highest utility of the latter for residential purposes. In the deposition given by Geary after argument the foregoing is essentially, with some differences in detail and emphasis, the approach he took in explaining why he used the sale in testifying in the *553Kaupelis case but not in this one. If Geary omitted the Mapland sale in his testimony in the present case for the dishonest purpose of keeping from the jury the relatively high value reflected by the sale, it is difficult to understand why he would not have had the same motive to omit the sale in testifying for the State in the Kaupelis case a few days later, where it was also to the State’s interest to show lower location value. Yet he did cite it there.
The question here is not the soundness of Geary’s approach or the probative weight of Ms testimony. That was for evaluation by the jury, which had the Mapland sale details before it, furnished, as noted above, by one of the defendant’s experts. The verdict has not been held to be against the weight of the evidence. The issue is solely whether what the trial court had before it, including the transcript of the Kaupelis trial, justified a nullification of the instant verdict on the basis of a finding that Geary gave willfully false testimony in the present case or omitted the Mapland sale in his testimony dishonestly. We are firmly of the view that no such conclusion should have been drawn — much less that such alleged willful falsity or dishonesty was shown “by clear, convincing and satisfactory evidence” or that Geary’s testimony as to the sale in either its positive or negative aspect could not reasonably have been met by defendant at the trial and that it “probably * * * controlled the result”. Shammas v. Shammas, supra (9 N. J. at 330).
We are not in accord with the view of our dissenting colleague that the matter should be remanded for redetermination by the trial judge as to the merits of the motion for new trial so far as based on the weight of the evidence. The judge’s conclusions do not suggest a view by him that if Geary’s testimony were not impugnable as willfuly false the verdict would be against the weight of the evidence. He could well have adopted that as an alternative ground for granting a new trial had that been his opinion but he did not. The verdict was, in our opinion, well within the evidence. Indeed, *554as to the large Probasco tract, as noted above, no witness on either side used the Mapland sale.
Reversed.