The opinion of the court was delivered by KING, P.J.A.D.

This is a taxpayer's appeal from a real property tax assessment involving the years 1981 and 1982. The taxpayer attacks the evaluation on several grounds. We reject plaintiff-appellant's contentions and affirm for the reasons given by Judge Crabtree in his written opinion of September 19, 1988.

On November 25, 1981 appellant filed a petition of appeal with the Bergen County Board of Taxation contesting the assessment of its parcel of land, Block 3712, Lot 5, Township of Teaneck, for the tax year 1981. On December 28, 1981 the Board of Taxation affirmed the assessment. On December 31, 1981 appellant filed a complaint in the Tax Court appealing the judgment of the Board of Taxation. On July 26, 1982 appellant filed a direct appeal with the Tax Court contesting the assessment of the same parcel for the tax year 1982.

Both matters were heard by Judge Crabtree in March and April 1988. On September 19th he rendered a written opinion and on October 6, 1988 judgment was entered in accordance with his opinion. The assessments were summarized as follows:1981 Original County Tax Board Tax Court Assessment Judgment Judgment 1982

The appellant raises four issues:

I. DID THE TAX COURT ERR IN FAILING TO PROVIDE AN ADEQUATE BASIS FOR ITS CONCLUSION AS TO VALUE FOR THE SUBJECT PROPERTY FOR THE TAX YEARS IN QUESTION?

II. DID THE TAX COURT ERR IN ACCEPTING THE COMPARABLES SET FORTH BY DEFENDANT'S EXPERT?

III. DID THE TAX COURT INCORRECTLY APPLY THE RATIO OF ASSESSED TO TRUE VALUE PURSUANT TO CHAPTER 123?

IV. DID THE TAX COURT ERR IN NOT ALLOWING PLAINTIFF'S COUNSEL TO CROSS-EXAMINE DEFENDANT'S EXPERT CONCERNING A CHART IN THE EXPERT'S REPORT DEALING WITH CHAPTER 123 RATIOS?

This is the factual background. The property in question was originally vacant land, designated as Block 3712, Lot 5, located on Home Street in the Township of Teaneck. The parcel comprises approximately 47.5 acres, including Block 4403, Lot 1, which is a separate tax parcel and which is not part of this appeal. The property has frontage on DeGraw Avenue, Glenwood Avenue and Marion Street, and is low (six to ten feet below grade), wet, irregularly shaped marshland. The property was purchased by appellant in March 1981 from the Teaneck Redevelopment Agency at a structured public sale for a total consideration of $1,052,910.77. The contract originally was dated September 12, 1977 and provided for an original purchase price of $1,000,000. The reason for the delay of the final sale of the property was a lawsuit filed by a citizens' group alleging that the property was worth considerably more than the purchase price. The site was acquired for a planned development to include townhouse, office, retail and hotel uses. At the time of sale the site was zoned for 567,000 square feet of office space, a 350-room hotel, 171 townhouses, and additional future condominium development. During 1981 the site was filled and prepared for development.

Michael Hiller, appellant's expert, said that the fair market value of the subject property as of the assessment date for the 1981 tax year, October 1, 1980, was the actual acquisition cost of March 1981, $1,052,910.77. From the purchase price Hiller deducted the value of Block 4403, Lot 1, for a total property value of $987,200. He prorated this amount for nine months, since the property for the first three months was owned by the Teaneck Redevelopment Agency and was tax exempt. Thus, Hiller found the true value to be $740,400 or $20,000 per acre. Hiller made no adjustment for the period between the execution of the contract and the closing date because such adjustments were offset by the rise in interest rates, cost of construction and cost of site improvements. Hiller also said that the property's highest and best use is its present use.

For corroborative purposes, Hiller used five comparables ranging in price per acre from $58,140 to $87,805. However, he primarily used these comparables for the 1982 tax year. In fact, Hiller conceded that only two of the sales were comparable and that the site work required on the subject property was considerably greater than that required for the other two sales. Further, Hiller was unaware of the allowable building coverage for any of the comparables he used. For the 1982 tax year, Hiller arrived at a total site value of $4,715,200, or $100,000 per acre. Hiller based that value on the acquisition cost of the land plus the site preparation costs.

Professor Arthur Cohen, Chairman of the Department of Statistics at Rutgers University, an expert in the field of mathematics and statistics, also testified on behalf of appellant. Cohen prepared three reports relating to the assessment practices of the Township of Teaneck for the years 1981-1987. Cohen performed his study to estimate the mean level of assessment-to-sales ratios in a given year based on unweighted, unclassified readings, contrary to the Chapter 123 ratios. See N.J.S.A. 54:1-35a. Dr. Cohen noted that his comparison between the Chapter 123 calculations and his computations showed a large percentage difference. Cohen said that he used the unweighted, unclassified ratios because they produced an accurate estimate of the mean level of assessment-to-sales ratio in a population in a given year. He noted that the discrepancies between the Chapter 123 ratios and the unweighted, unclassified ratios in cases such as the present one are so dramatic that "Chapter 123 is totally misleading as an assessment of the average level . . ." Cohen criticized the Chapter 123 methodology on the grounds that it uses total instead of individual ratios and because it uses data from previous years.

Respondent's expert, John Lasser, relied exclusively on the market approach to value for the 1981 tax year. He stated that the best and most profitable use of the property was its current multi-purpose use. He used six comparables, ranging in price per acre from about $60,000 to about $137,000, and ranging in size from about 6.5 acres to over 19 acres. The comparable sales took place between August 1979 and September 1980. The sales were in nearby communities in Bergen County. Most of the comparables were low, wet marshland similar to the subject property. Based on these sales, Lasser testified that the true value of the appellant's land was $100,000 per acre for the 1981 tax year — $4,736,100 for the entire property. Lasser did not rely on appellant's contract of purchase price since the contract had been made a "substantial period" before the closing. For the 1982 tax year Lasser used the same comparables, resulting in a true value of $110,000 per acre. To that figure Lasser added $3,311,600 for fill, resulting in a total for land of $8,521,300. Lasser estimated that the improvements for the 1982 tax year — concrete footings and foundations for various buildings — amounted to $1,659,000. Therefore, according to Lasser, the total value for the 1982 tax year was $10,180,300.

For the 1981 tax year, Judge Crabtree found that the contract price urged by appellant's expert was not probative of true value because the contract had been executed three-and-one-half years before the assessing date; this, he found, was too remote in time to provide a reliable indication of value. He also found respondent's expert more persuasive because he chose properties comparable in physical characteristics, location and date of sale. However, he did not agree with the expert's value estimate of $100,000 per acre and reduced the true value to $80,000 per acre. Accordingly, he found the true value of the subject property on October 1, 1980 to be $3,800,000. He further found that the ratio of assessment to true value was 31.25%. Since the ratio fell below the lower limit of the common level range, 72% (the average ratio for the 1981 tax year was 85%), he increased the assessment by applying the average ratio of 85% to the true value, resulting in an annual assessment of $3,230,000. Since the property was tax exempt for the first three months of the year, the assessment was prorated for nine months, resulting in a final assessment of $2,422,500.

For the 1982 tax year, Judge Crabtree rejected appellant's valuation for the same reasons as he did for the 1981 tax year. He found respondent's expert's estimate persuasive to the extent that it was based on a 10% increase in property values from 1981-1982, an increase which the judge called "conservative." Therefore, the judge found the true value of the land to be $88,000 per acre as of October 1, 1981. Further, the judge held that profit earned by appellant as its own general contractor, in the amount of 10% of actual costs, should be added to the direct costs. Accordingly, he added costs for site improvements, townhouse improvements, east office building improvements and retail shops improvements to the value of the land, resulting in a true value of $15,361,200. In turn, the judge applied the Director's 1982 Chapter 123 ratio of 77% to increase the assessment to $11,828,000 because the ratio of assessment to true value was 39.73% which fell below the lower limit of the common level range, 65%.

I

Appellant contends that the Tax Court Judge did not justify or explain how he determined the true value of the property and certain cost figures for the tax years in question. Respondent argues that the judge was fully justified both in rejecting the sale of the subject property, standing alone, as evidence of market value, and in adopting the comparable sales approach as the true indication of value.

As to the 1981 tax year, Judge Crabtree ruled as follows:

I cannot accept the opinion of value proffered by plaintiff's expert. The contract purchase price is not probative of true value as the contract was executed 3 1/2 years before the assessing date, which is too remote in time to provide a reliable indication of value. The expert's three-fold explanation for disregarding the time interval is not convincing. He admitted that increased costs and interest rates affected all vacant land in Bergen County; yet he made no study of vacant land sales to prove the absence of appreciation in land values during the time interval in question.

The evidence of defendant's expert is more persuasive. He chose properties comparable in physical characteristics, proximate in location and sales reasonably close to the assessing date.

As to the 1982 tax year, he stated:

Plaintiff's expert's land valuation, which, like his estimate for tax year 1981, is based upon the land acquisition price set out in the 1977 agreement, is rejected for the reasons heretofore given, i.e., the 1977 contract price is not probative of the October 1, 1980 value. Its probative force is even less one year later. Defendant's expert's estimate, on the other hand, is persuasive, at least to the extent that it is based upon a 10% increase in property values from 1981 to 1982. That magnitude of increase is, if anything, conservative. The average ratio of assessments to selling prices of Teaneck properties declined from 96% as of October 1, 1980 to 85% as of October 1, 1981, showing an 11.4% increase in property values in a 12-month period.

The judges presiding in the Tax Court have special expertise; for that reason their findings will not be disturbed unless they are plainly arbitrary or there is a lack of substantial evidence to support them. Southbridge Park, Inc. v. Borough of Fort Lee, 201 N.J. Super. 91 , 94, 492 A.2d 1026 (App.Div. 1985); G S Co. v. Eatontown Bor., 6 N.J. Tax 218 , 220 (App.Div. 1982). See N.J.S.A. 2A:3A-13. The Tax Court has the duty to apply its own judgment to valuation data submitted by experts in order to arrive at true value. Warren Tp. v. Suffness, 225 N.J. Super. 399, 414, 542 A.2d 931 (App.Div. 1988), certif. den. 113 N.J. 640, 552 A.2d 166 (1988). However, the Tax Court's right to make an independent assessment is not boundless; it must be based on evidence before it and data that are properly at its disposal, F.M.C. Stores Co. v. Borough of Morris Plains, 100 N.J. 418 , 430, 495 A.2d 1313 (1985). It must not arbitrarily assign a value to the property which is not supported in the record. Warren Tp. v. Suffness, 225 N.J. Super. at 414, 542 A.2d 931.

The essence of appellant's challenge in this instance is that Judge Crabtree's findings are ungrounded in substantial credible evidence. The judge rejected the opinion of value offered by appellant's expert because the contract purchase price was "too remote in time to provide a reliable indication of value." He found the opinion of respondent's expert, based on the comparable sales approach, "more persuasive," but determined that the true value should be somewhat less than urged by respondent — $80,000 per acres for 1981, and $88,000 for 1982. With respect to the improvements, for the 1982 tax year he found the original cost of construction "controlling." In addition, he held that a contractor's profit must be added to these costs when, as here, a party acts as its own general contractor. He found that the cost figures relied on by the parties were "substantially lower than the actual cost figures disclosed by the credible evidence." Therefore, Judge Crabtree found total true value for the 1982 tax year to be $15,361,200.

Based on these facts, we find no merit to appellant's argument that the judge did not justify or explain how he arrived at the figures to reflect true value. The judge's findings refute that argument. Appellant also claims that the judge did not set forth the basis for his conclusion that the cost figures were greater than that relied on by the parties. However, the Tax Court judge has the right to apply his own judgment to valuation data submitted by experts in order to arrive at true value. Glen Wall Associates v. Wall Tp., 99 N.J. 265 , 280, 491 A.2d 1247 (1985). Moreover, in resolving issues of fact arising from evidence given by such experts, the Tax Court is free to reject in whole or in part the testimony of such experts. Southbridge Park, Inc. v. Borough of Fort Lee, 201 N.J. Super. at 94, 492 A.2d 1026.

II

Appellant alleges that the comparables relied on by the judge, as set forth by respondent's expert, were in fact not comparable because of various differences including size, topography, location, off-site development, zoning restriction and tax rate. Concerning Lasser's comparable sales comparisons, Judge Crabtree stated:

The evidence of defendant's expert is more persuasive. He chose properties comparable in physical characteristics, proximate in location and sales reasonably close to the assessing date.

Under the market or comparable sales approach to valuation, comparable properties are analyzed and adjusted for variables. American Institute of Real Estate Appraisers, The Appraisal of Real Estate, (9th ed. 1987) at 69. Evidence of comparable sales is effective in determining value only where there is a substantial similarity between the properties to admit of reasonable comparison. Venino v. Borough of Carlstadt, 1 N.J. Tax 172 , 175 (Tax Ct. 1980), aff'd 4 N.J. Tax 528 (App.Div. 198 1).

Lasser used six comparable sales from nearby communities in Bergen County, all of which occurred within about a year of the assessment date. The sales ranged in size from about 6.5 acres to over 19 acres, and in price per acre from approximately $59,500 to $137,000. Three of the sales were of unimproved land. Lasser conceded certain differences between the subject property and his first comparable, including the fact that the subject was some three times greater in size. He also conceded that he did not make any adjustments to this comparable for the off-site improvements required on the subject property. Despite the fact that his second comparable was only approximately 20% the size of the subject, Lasser stated that it was comparable because it had a similar use and was located in a similar area as the subject. Lasser acknowledged that he did not know the permitted building coverage on comparables three and six. He also testified that comparables four and five were located in a highly developed commercial and industrial park, while the subject was not.

Appellant correctly points to several differences between the comparables and the subject property. However, in the market approach, differences between a comparable property and the subject property are anticipated; they are dealt with by adjustments which recognize and explain the differences and then by relating the two properties to each other in a meaningful way so that an estimate of value can be determined. U.S. Life Realty Corp. v. Jackson Tp., 9 N.J. Tax 66 , 72 (Tax Ct. 1987). Here, Judge Crabtree found the comparables similar in physical characteristics, location and time of sale but recognized the differences as shown by his adjustment of the value estimate down from $100,000 per acre to $80,000 per acre.

This case is unlike Schmertz v. Dover Tp., 4 N.J. Tax 145 (Tax Ct. 1982), where the expert admitted that the comparables he relied on gave him "a lot of problems" and the judge found that the number of adjustments and the distances of the comparable sales from the subject rendered the comparables "meaningless" for purposes of establishing the value of the subject property. Nor is this case similar to Center-Whiteman Corp. v. Fort Lee, 4 N.J. Tax 153 (Tax Ct. 1982), where the comparisons between the subject and the comparables were found to be "rudimentary," where no testimony was offered as to their similarities, and where the comparable sales took place at least two and one-half years before the assessment date. Also distinguishable is Halocarbon Products v. Boro. of South River, 1 N.J. Tax 294 (Tax Ct. 1980), aff'd o.b. 181 N.J. Super. 1, 436 A.2d 532 (App.Div. 198 1), where the three comparables were "substantially smaller" than the subject (2.5, 5.25 and 8.4 acres as compared to nearly 38 acres) and one comparable involved a purchaser who was under economic compulsion to purchase the property. Another comparable was not affected by the flood hazard regulations which affected the subject property. In the context of comparables only about 10% the size of a subject property, the Tax Court has said that the "vast difference" in the size of the tracts may diminish the weight to be given in determining the value of the larger tract, but size differences alone do not disqualify them from consideration. Linwood Properties, Inc. v. Fort Lee, 7 N.J. Tax 320 , 336 (Tax Ct. 1985). Appellant also points to differences in topography between the subject property and the comparables. However, a difference in topography does not disqualify a comparable. See Owens-Illinois Glass Co. v. Bridgeton, 8 N.J. Tax 495 (Tax Ct. 1986). We do not believe that the differences here between the comparables and the subject are so great as to warrant a legal finding that the market approach should be rejected.

Moreover, appellant's alternative measure, the sales price of the subject property, is of little aid in determining value. While the sales price of the subject property is evidential, it is not the exclusive criterion of true value. Rek Investment Co. v. Newark, 80 N.J. Super. 552 , 559, 194 A.2d 368 (App.Div. 196 3). "It has been repeatedly held that special circumstances may increase or depress market value without affecting the true value of property." Additionally, the sale of the subject took place three and one-half years before the assessment date, too remote in time to provide a reliable indication of value.

III

Appellant contends that the Tax Court erred in applying the Chapter 123 ratios to the subject property where appellant had presented competent and uncontroverted testimony by Dr. Cohen that application of the Chapter 123 ratios was "egregious." Appellant contends that the Tax Court should have applied Cohen's unweighted and unclassified ratios. Respondent contends that appellant's argument amounts to no more than an attack on the methodology adopted by the Legislature and approved by the courts, which balances the desires of taxpayers for fairness on the one hand and the need for stability in assessments on the other.

On this issue, Judge Crabtree held:

Plaintiff offered the testimony of Dr. Arthur Cohen, a statistician, to show that ch. 123 provided an inadequate remedy for assessment discrimination relief. From a statistical sampling of 291 sales, Dr. Cohen found the average assessment-to-sales ratio to be 65.24%. The differences between the two ratios forecloses meaningful comparison. Dr. Cohen's ratio is unweighted and unclassified and covers sales for only one year, whereas the ch. 123 ratio is weighted by property class and by year and covers several years. Dr. Cohen's ratio is derived from sales occurring in the 12 months ended June 30, 1982; ch. 123 is determined as of October 1, 1981, partly on the basis of sales occurring during the 12 months ended June 30, 1981. As a result of these disparities plaintiff is unable to show that the application of ch. 123 leaves it with a confiscatory assessment, which is required by our Supreme Court in order for a taxpayer to obtain relief beyond ch. 123. Murnick v. Asbury Park, 95 N.J. 452 , 471 A.2d 1196 (1984).

The New Jersey Constitution requires that all property within a taxing district be assessed in accordance with the same standard of value. N.J. Const. (1947), Art. VIII, § 1, par. 1. Axiomatically in New Jersey, where a taxpayer's property is assessed at a ratio to true value substantially higher than the common level of assessments or, if there is no such common level substantially higher than the director's average ratio, that taxpayer is ordinarily entitled to a reduction of his assessment down to the common level or Director's ratio. Kearny Leasing Corp. v. Kearny, 6 N.J. Tax 363 , 378 (Tax Ct. 1984), aff'd 7 N.J. Tax 665 (App.Div. 1985), certif. den. 102 N.J. 340, 508 A.2d 215 (1985). N.J.S.A. 54:51A-6, commonly known as the Chapter 123 ratio, provides in pertinent part:

a. Whenever the tax court is satisfied by the proofs that the ratio of the assessed valuation of the subject property to its true value exceeds the upper limit or falls below the lower limit of the common level range, it shall enter judgment revising the taxable value of the property by applying the average ratio to the true value of the property except as hereinafter provided.

b. If the average ratio is below the county percentage level and the ratio of the assessed value of the subject property to its true value exceeds the county percentage level, the tax court shall enter judgment revising the taxable value of the property by applying the average ratio to the true value of the property.

c. If both the average ratio and the ratio of the assessed value of the subject property to its true value exceed the county percentage level, the tax court shall enter judgment revising the taxable value of the property by applying the county percentage level to the true value of the property.

The average ratio of assessment to true value of real property for a taxing district is the ratio promulgated by the Director of the Division of Taxation as of October 1 of the year preceding the tax year in question, as revised by the Tax Court. N.J.S.A. 54:1-35a. The "common level range" for a taxing district is that range which is plus or minus 15% of the average ratio for that district. N.J.S.A. 54:1-35b. In determining the average ratio, the Director considers sales and assessed values within the taxing district, exclusive of twenty-seven categories of deed transactions. N.J.A.C. 18:12-1.1(a). As initially enacted, the average ratio for Chapter 123 was "the unweighted unclassified arithmetic average as of October 1 of the year preceding the tax year. . . ." L. 1973, c. 123, § 1(a). The definition of the average ratio in Chapter 123 was changed in 1979 to a weighted, classified average. L. 1979, c. 114, § 10.

Chapter 123 establishes a rebuttable presumption of the common level assessment. Murnick v. Asbury Park, 95 N.J. 452 , 462, 471 A.2d 1196 (1984). The underlying purpose of the Act was to relax the taxpayer's burden of proof. Id. at 462-463, 471 A.2d 1196. In the absence of additional proof, a taxpayer may not substitute another ratio, and to overcome the presumption, the taxpayer must establish that application of the ratio would be "virtually confiscatory." Id. at 463, 471 A.2d 1196. "As a practical matter, the presumption created by chapter 123 is so strong that it will be conclusive in all but the most egregious cases." Ibid.

Dr. Cohen said in his testimony that he used unweighted and unclassified averages. In Murnick, the Court discussed these terms:

The terms "unweighted" and "unclassified," like "weighted" and "classified," are not defined in either the relevant statutes or administrative regulations. Speaking generally, an unweighted ratio weighs the assessment ratio of each parcel equally, but a weighted ratio gives greater weight to the assessment of parcels that are sold for higher prices. An unclassified ratio does not distinguish between different kinds of property, but a classified ratio categorizes property by group as, for example, vacant, residential, farm or other (including commercial and industrial) property and calculates the ratio on that basis. [ 95 N.J. at 461, 471 A.2d 1196.]

According to Cohen, there was a large percentage difference between the Chapter 123 ratio and the unweighted, unclassified ratio which he developed. He stated that, based on his expertise as a statistician, the discrepancies between the two ratios were so "dramatic" that "Chapter 123 is totally misleading as an assessment at the average level." Cohen further opined that the Chapter 123 ratio was "not even in the ballpark" of measuring the "central tendency" of the Teaneck population in a given year. Cohen concluded that as an estimate of the mean of the population in a given year, the Chapter 123 ratio was highly inaccurate and heavily biased upwards, while his method gave an accurate estimate. Therefore, while the Director's ratio, promulgated under Chapter 123, was 85% for 1981 and 77% for 1982, Cohen's ratio was 69.28% for 1981 and 65.24% for 1982.

In our view Rudd v. Cranford Tp., 4 N.J. Tax 236 (Tax Ct. 1982), refutes appellant's argument. In Rudd, the plaintiff asserted that the relief from assessment discrimination provided by statute was inadequate, and that sufficient relief could only be afforded by applying an unweighted, unclassified ratio to the true value of the property. Id. at 247. In support of his contention, the plaintiff attempted to show that the unweighted, unclassified ratio was substantially lower than the Chapter 123 ratio. Ibid. The Tax Court rejected the plaintiff's contention, holding that the unweighted ratio "is subject to temporary market fluctuations and ascribes inordinate influence to inadequate sales samples in a single year. The Legislature recognized this when it changed the chapter 123 ratio from a one-year, unweighted and unclassified ratio to the classified and weighted ratio . . ." Id. at 249. The Tax Court also noted that the plaintiff used a ratio promulgated as of October 1 of the tax year, rather than October 1 of the pretax year as required by Chapter 123. Id. at 248. The court described this comparison as "inappropriate" because of "the obvious fact of a significant annual appreciation in value of the generality of properties in [the] municipality . . ." Id. at 248-249.

Similarly, in this case appellant's attempt to substitute the unweighted, unclassified ratio for the Chapter 123 ratio must be rejected. As noted, the unweighted, unclassified ratio was discarded by the Legislature in 1979. As observed in Murnick, the unweighted, unclassified ratio does not distinguish between different types of property. 95 N.J. at 461, 471 A.2d 1196. Moreover, as pointed out in Rudd, this kind of ratio is subject to temporary market fluctuations and the inordinate influence of inadequate sales samples. 4 N.J. Tax at 249. In our view, the differences in the ratios for the two years in issue are not "virtually confiscatory," as required by Murnick. In addition, appellant made the same error in this case as did the plaintiff in Rudd, namely, relying on the ratio promulgated in the tax year in question rather than the year preceding the tax year. As concluded in Rudd, this is "inappropriate." 4 N.J. Tax at 248-249. Reliance on Cohen's intricate statistics in itself would appear to contravene the Legislature's motivation for enacting Chapter 123, since prior to enactment "a party [had] labored to prove its case by extensive statistical studies introduced through experts." Kearny Leasing Corp. v. Kearny, supra, 6 N.J. Tax at 378. Finally, as the Supreme Court noted in Murnick, "a taxpayer's right to relief should be determined in accordance with chapter 123 in all but the most extreme or severe circumstances." 95 N.J. at 463, 471 A.2d 1196. We do not find that this is such a circumstance.

IV

We conclude that appellant's final point on the restriction of its cross-examination of respondent's expert, is clearly without merit. R. 2:11-3(e)(1)(E). While an expert witness is always subject to searching cross-examination as to the basis of his opinion, County of Ocean v. Landolfo, 132 N.J. Super. 523 , 528, 334 A.2d 360 (App.Div. 1975), ordinarily, the scope of cross-examination of a witness rests in the discretion of the trial judge. An appellate court will not interfere with the exercise of such discretion unless clear error and prejudice are shown. Janus v. Hackensack Hospital, 131 N.J. Super. 535 , 540, 330 A.2d 628 (App.Div. 1974), certif. den. 67 N.J. 95, 335 A.2d 47 (1975). Further, the extent of allowable cross-examination where it involves collateral matters is measured by the trial judge in light of the effect such examination may have upon substantial justice. Mazza v. Winters, 95 N.J. Super. 71 , 78, 230 A.2d 139 (App.Div. 196 7). Moreover, cross-examination is usually limited to matters discussed on direct examination. State v. Mustacchio, 109 N.J. Super. 257 , 264, 263 A.2d 139 (App.Div. 1970), aff'd 57 N.J. 265, 271 A.2d 582 (1970).

Affirmed.