Defendant Virginia Chemicals appeals in a personal injury products liability action. Plaintiff Frank Halloran, an automobile mechanic, obtained a verdict in his favor, after a jury trial on the issue of liability only, for injuries he sustained while using a can of refrigerant packaged and sold by the chemical company. A divided Appellate Division affirmed, and certified a question of law for review in this court. *

The principal issue argued by defendant Virginia Chemicals is that plaintiff failed to make out a prima facie case because no particular defect in the packaged refrigerant was ever discovered or established. That issue merits little discussion. In a products liability case it is now established that, if plaintiff has proven that the product has not performed as intended and excluded all causes of the accident not attributable to defendant, the fact finder may, even if the particular defect has not been proven, infer that the accident could only have occurred due to some defect in the product or its packaging (see Codling v Paglia, 32 NY2d 330, 337-338; Fogal v Genesee Hosp., 41 AD2d 468, 478; see, generally, 47 NY Jur, Products Liability, §§ 14, 18).

There is one other issue meriting extended discussion: whether evidence that the injured mechanic had previously used an immersion heating coil to heat the can of the refrigerant should be admissible to show that on the particular occasion he was negligent and ignored the labeled warnings on the can. Evidently relying on the rubric excluding prior[*389] instances of carelessness to create an inference of carelessness on a particular occasion, both the Trial Judge and the Appellate Division, save for two dissenting Justices, agreed that such evidence was not admissible.

There should be a reversal and a new trial. If plaintiff, when necessary to stimulate the flow of the refrigerant, a highly compressed liquefied gas, habitually or regularly used an immersion coil to heat the water in which the container was placed, evidence of that habit or regular usage should be admissible to prove he followed such a procedure on the day of the explosion. Evidence of habit or regular usage, if properly defined and therefore circumscribed, involves more than unpatterned occasional conduct, that is, conduct however frequent yet likely to vary from time to time depending upon the surrounding circumstances; it involves a repetitive pattern of conduct and therefore predictable and predictive conduct. On this view, the excluded evidence was offered to show a particular method of executing a task followed by the mechanic, who, on his own testimony, had serviced "hundreds” of air-conditioning units and used "thousands” of cans of the refrigerant. If on remittal the evidence tends to show that the mechanic used an immersion coil a sufficient number of times to warrant a finding of habit, or regular usage, it would be admissible to aid the jury on its inquiry whether he did so on the occasion in question.

On June 1, 1970, the day of the accident, Frank Halloran, a mechanic for 15 years, had been employed by the Hillcrest Service Station for over three years. Among his duties was the servicing and charging of automobile air-conditioning units, a job for which he had been specially trained, and for which he used "all [his] own tools.” The particular task involved that day was the changing of the air-conditioning compressor on a 1967 Chrysler automobile. Plaintiff testified that he had emptied the system, removed the old compressor, and installed a new one. He then began to charge the unit.

The first two cans of the refrigerant, Freon, flowed into the system without difficulty. By the time he was emptying the third can, however, plaintiff found it necessary to accelerate the flow of the refrigerant. The mechanic described how he filled an empty two-pound coffee tin with warm tap water, used a thermometer to determine that the water temperature was about 90 to 100 degrees, and inserted into the coffee tin the third can of Freon. Having a similar problem with the[*390] flow of the fourth can, Halloran again dropped the Freon into the warm water. Noticing that his low pressure gauge showed a rapid increase in the pressure, and aware that "something was wrong”, Halloran reached down to remove the can from the water, but was too late. The can exploded before he could touch it.

Neither the thermometer Halloran claimed to have used nor the bottom of the exploded can of Freon was produced at trial. Halloran knew that excessive heating of the can would cause damage, and that the warnings on the can specified 130 degrees as the maximum permissible safe temperature. As discussed earlier, he proved no particular defect in the can, its contents, or in so much of the exploded can which was produced at the trial. Having worked alone that day, Halloran was the only eyewitness to the explosion.

Defendant Virginia Chemicals, on cross-examination of Halloran and on its defense, sought to establish that it was Halloran’s "usage and practice” to use an immersion coil to heat the water in which the Freon was placed. Halloran denied ever making such use of an immersion coil. But defendant offered a witness prepared to testify not only that he had seen Halloran on previous occasions using an immersion coil to heat Freon, but that he had warned plaintiff of the danger as well. Plaintiff, relying on the rule that extrinsic evidence cannot be introduced to impeach a witness on collateral matters, objected to the admissibility of such testimony. The Trial Judge sustained the objection.

Of course, had an immersion heating coil been used at the time of the accident the unexplained and thus far unexplainable explosion would have been fully explained.

Were the evidence defendant sought to produce collateral, defendant generally would be bound by plaintiff’s denial. For it is now well settled that extrinsic evidence introduced solely to impeach credibility on a collateral issue is, with special exceptions, inadmissible (People v Schwartzman, 24 NY2d 241, 245; Potter v Browne, 197 NY 288, 293; Richardson, Evidence [10th ed], § 491).

To be sure, Halloran’s practice prior to June 1, 1970 is not conclusive proof of the method he employed in working on the 1967 Chrysler. "Collateral”, however, it is not. Logically probative it is and ought to be. While courts of this State have in negligence cases traditionally excluded evidence of carefulness or carelessness as not probative of how one acted on a particu[*391] lar occasion, in other cases evidence of a consistent practice or method followed by a person has routinely been allowed (compare, e.g., Zucker v Whitridge, 205 NY 50, 58-66 [carefulness], and Hartley v Szadkowski, 32 AD2d 550 [carelessness], with People v Bombard, 5 AD2d 923, cert den 358 US 849 [prosecutor’s practice of insisting defendant be advised of right to counsel]). That a kind of habit, practice, or method was proffered in this case to establish negligence should not, without more, affect its admissibility.

Because one who has demonstrated a consistent response under given circumstances is more likely to repeat that response when the circumstances arise again, evidence of habit has, since the days of the common-law reports, generally been admissible to prove conformity on specified occasions (e.g., Miller v Hackley, 5 Johns 375, 384; 1 Wigmore, Evidence [3d ed], § 92; Richardson, Evidence [10th ed], § 185; Fed Rules Evidence, rule 406 [in US Code, tit 28, Appendix]; see, generally, Lewan, Rationale of Habit Evidence, 16 Syracuse L Rev 39). Hence, a lawyer, to prove due execution of a will, may testify that he always has wills executed according to statutory requirements (Matter of Kellum, 52 NY 517, 519-520). So too, to prove that notice is mailed on a specified day of the month, one is allowed to testify that he is in the habit of being home on that day of the month to transact such business (Beakes v Da Cunha, 126 NY 293, 298).

When negligence is at issue, however, New York courts have long resisted allowing evidence of specific acts of carelessness or carefulness to create an inference that such conduct was repeated when like circumstances were again presented (e.g., Warner v New York Cent. R. R. Co., 44 NY 465, 472; Grenadier v Surface Transp. Corp. of N. Y., 271 App Div 460, 461; Lefcourt v Jenkinson, 258 App Div 1080; Richardson, Evidence, § 186; Fisch, New York Evidence, § 202; cf. Cabezudo v New York’s Eldorado, 50 AD2d 794, 795). Hence, evidence of a plaintiff’s habit of jumping on streetcars may not be offered to prove he was negligent on the day of the accident (Eppendorf v Brooklyn City & Newtown R. R. Co., 69 NY 195, 197). Nor could testimony that the deceased had usually looked both ways before crossing railroad tracks be introduced to establish his care on the particular occasion (Zucker v Whitridge 205 NY 50, 58-66, supra). Whether a carry-over from the prohibition against using so-called "character” evidence in civil cases, or grounded on the assumption that even repeated[*392] instances of negligence or care do not sufficiently increase the probability of like conduct on a particular occasion, the statement that evidence of habit or regular usage is never admissible to establish negligence is too broad (see 1 Wigmore, Evidence [3d ed], § 97, esp p 532).

At least, as in this kind of case, where the issue involves proof of a deliberate and repetitive practice, a party should be able, by introducing evidence of such habit or regular usage, to allow the inference of its persistence, and hence negligence on a particular occasion (see McCormick, Evidence [2d ed], § 195, advocating an even more expansive approach; see, also, 1 Wigmore, Evidence [3d ed], § 97). Far less likely to vary with the attendant circumstances, such repetitive conduct is more predictive than the frequency (or rarity) of jumping on streetcars or exercising stop-look-and-listen caution in crossing railroad tracks. On no view, under traditional analysis, can conduct involving not only oneself but particularly other persons or independently controlled instrumentalities produce a regular usage because of the likely variation of the circumstances in which such conduct will be indulged. Proof of a deliberate repetitive practice by one in complete control of the circumstances is quite another matter and it should therefore be admissible because it is so highly probative.

As previously noted, Halloran, in the course of his work as a mechanic, had serviced "hundreds” of automobile air conditioners and had used "thousands” of cans of Freon. From his testimony at trial it seems clear that in servicing these units he followed, as of course he would, a routine. If, indeed, the use of an immersion coil tended to be part of this routine whenever it was necessary to accelerate the flow of the refrigerant, as he indicated was often the case, the jury should not be precluded from considering such evidence as an aid to its determination.

Of course, to justify introduction of habit or regular usage, a party must be able to show on voir dire, to the satisfaction of the Trial Judge, that he expects to prove a sufficient number of instances of the conduct in question (see Fed Rules Evidence, rule 406; Model Code of Evidence, rule 307, subd [3], par [b], including Comment, Illustrations 2, 3 [1942]; Uniform Rules of Evidence, rule 406, subd [b]; see, also, 2 Wigmore, Evidence [3d ed], § 376). If defendant’s witness was prepared to testify to seeing Halloran using an immersion coil on only one occasion, exclusion was proper. If, on the other hand, plaintiff[*393] was seen a sufficient number of times, and it is preferable that defendant be able to fix, at least generally, the times and places of such occurrences, a finding of habit or regular usage would be warranted and the evidence admissible for the jury’s consideration.

Plaintiff in his brief raised no procedural issue, but the obvious question of whether the chemical company preserved its objection to the ruling excluding its proffer of evidence of Halloran’s prior practice in using an immersion coil is not avoidable (see CPLR 5501, subd [a], par 3; 4017).

Defendant offered the testimony to prove Halloran’s "usage and practice”. Plaintiff objected that it was collateral, and defendant responded that, since plaintiff denied ever using an immersion coil, the evidence was admissible on the issue of plaintiff’s credibility. The court in excluding the proffered evidence did not give its reasoning, but evidently adopted plaintiff’s contention that the evidence related only to a collateral issue.

True, it was on cross-examination that Halloran testified he never used an immersion coil in heating cans of Freon. And, as previously noted, were the issue truly collateral, defendant indeed would, in the great majority of situations, be precluded from introducing extrinsic evidence to impeach plaintiff’s credibility on such a matter (e.g., Potter v Browne, 197 NY 288, 293, supra). On direct examination, however, Halloran testified to his regular practice in using Freon. It involved no more than using warm tap water in the coffee tin, checking with a thermometer to prevent excessive heat, and immersing the Freon can in the warm water. He always read, repetitiously, before any use of the refrigerant, the warning instructions on the label about maximum safe temperatures.

Consequently, the practice he followed, even if it were not admissible as evidence of habit, became a material issue as to which contradictory evidence was not precluded, even if, as defendant mistakenly urged, it involved solely credibility. Halloran had "opened the door” on the issue, made what he later argued was a collateral issue a material one, and he could not benefit from his testimony on direct examination, yet bar its refutation on defendant’s case (see People v Leonardo, 199 NY 432, 441-442; Blossom v Barrett, 37 NY 434, 438, and cases cited). Defendant’s objection to the court’s ruling, therefore, adequately preserved the issue for review.[*394] Of course, on a new trial defendant would no longer be limited to its too narrow objection, if too narrow it was.

With respect to Virginia Chemicals’ third-party complaint, dismissal was recited in the judgment, but not in the ordering paragraphs (see order of affirmance at App Div 50 AD2d 852). The omission was a matter of form, correctable for the asking by a motion to resettle the interlocutory judgment (see CPLR 2001, 5019, subd [a]; 5 Weinstein-Korn-Miller, NY Civ Prac, par 5019.03; cf. Marino v Nolan, 29 AD2d 541, affd 21 NY2d 738). The judgment should be accorded, therefore, the effect patently intended (see CPLR 5019, subd [a]). Virginia Chemicals, however, offers no persuasive argument why the third-party complaint should be reinstated, and none is perceived.

Accordingly, the order of the Appellate Division should be modified by reversing so much of the order as affirmed the award of judgment to plaintiffs, on their complaint, and to A & E, on its cross claim against Virginia Chemicals, with costs to abide the event, and affirming as to the dismissal of the third-party complaint, intended by the decision and judgment of the Supreme Court, with costs to third-party defendants, and the action remitted for a new trial on the issue of liability.

Judges Jasen, Gabrielli, Jones, Wachtler, Fuchsberg and Cooke concur.

Order modified, and the case remitted to Supreme Court, Queens County, for a new trial in accordance with the opinion herein, with costs to abide the event, and, as so modified, affirmed, with costs to third-party defendants-respondents. Question certified answered in the negative.