OPINION OF THE COURT
Bertram Katz, J.It is a legal truism that harsh sanctions breed ingenuity. Thus, when the New York State Legislature stiffened the penalties against intoxicated drivers two years ago,1 one predictable result was that the number of challenges to the use of blood alcohol measuring devices would increase proportionately.2
One such example is the defendant in this case, Porfirio Molina, whose counsel has submitted a numerically im*484pressive list of challenges to the introduction of the breathalyzer test results, five in all. I have relegated a description of the operation of the breathalyzer machine to a footnote,3 since the subject has been extensively covered in other opinions.
Mr. Molina was arrested on December 2, 1982 and brought by the police to their testing station, called Highway 1, which is located on the Bronx River Parkway. This testing station, which serves The Bronx and Manhattan, is one of several throughout New York City. Mr. Molina’s breath, as analyzed by the Smith and Wesson model 900A breathalyzer, allegedly yielded a reading of .21%. He accordingly was charged with a violation of subdivision 2 of section 1192 of the Vehicle and Traffic Law, driving with a blood alcohol percentage exceeding .10%, which is a misdemeanor, but a predicate to a felony.
Subsequently, the defendant requested, as part of discovery, the breathalyzer ampoules used in his test for retesting and examination. As a routine practice, the police in New York City destroy the ampoules and their contents immediately after use, and do not preserve breath samples. The defendant has now moved to suppress the breathalyzer results on five grounds, four involving the ampoules, and one involving breath samples.
The defendant claims that destruction of the ampoules deprives him of due process of law in that he is thereby prevented from:
(1) Retesting the used ampoule to confirm or contradict the original blood alcohol reading, a practice recently *485found to be impossible in People v Santiago (116 Misc 2d 340);
(2) Re-examining the glass ampoule itself for any gross defects, such as size;
(3) Re-examining the ampoule’s chemical contents to determine whether the solution was proper in volume; and
(4) Re-examining the ampoule’s contents to determine whether the chemicals were mixed in their proper proportions.
It should be noted parenthetically that a suppression motion based on the latter three grounds was decided in favor of the defendants, over the People’s default in proof, by my colleague, Judge Greenberg, in People v Bascombe (NYLJ, Feb. 25, 1983, p 12, col 2).
Finally, the defendant would impose upon the police a duty to capture and preserve an additional breath sample, to be maintained for independent testing by the defendant. In the absence of this safeguard, the defendant argues, the admission of the breathalyzer test results is a denial of due process. (Garcia v District Ct., 197 Col 38.)
THE HEARING
In an extended hearing encompassing all five issues,4 the defendant relied on his one expert witness, chemist and industrial hygienist, Dr. Harvey Cohen, who has conducted experiments into the preservation of ampoules. By contrast, the People called five witnesses, including the police officers who arrested and tested Mr. Molina. The police lab technician described the police procedures used in maintaining the breathalyzer machines. A technician for the company that conducts quality control tests on the test ampoules, Stifel Research Institute, revealed the process by which ampoules are screened for defects. Finally, the People’s breathalyzer expert, Michigan police toxicologist Dr. Edgar Kivela, testified as to the consensus of scientific opinion on preservation of ampoules and breath samples. After having examined the Various exhibits, which included two types of breath-trapping paraphernalia *486and having heard hours of polished expert testimony, I make the following findings of fact.
FINDINGS
(1) Despite the interesting, though inconclusive experiments of Dr. Cohen,51 find that retesting of ampoules for blood alcohol, which is still more in the realm of theory than of established fact, has not attained wide acceptance in the scientific community. Accordingly, such testimony is inadmissible. (Frye v United States, 293 F 1013; State v Canaday, 90 Wn 2d 808; People v Leone, 25 NY2d 511; People v Hughes, 59 NY2d 523; Wigmore, Evidence [3d ed], § 990.)
(2) Re-examination of ampoules for gross defects, for insufficient chemical volume, or for defective fluid composition may be possible at a later date, and may yield some useful information to the defendant. However, there is no indication in the hearing transcript of how this preservation is to be accomplished. In addition, the burden on the police of saving thousands of such ampoules, both “test” and “control”, together with volatile chemicals, is not to be minimized. There was ample testimony by the representative of Stifel that quality control tests effectively weed out all defective ampoules before they reach the police. The most significant factor, however, was the testimony of Dr. Kivela and others, who described how the so-called “checklist” procedures of “purging”, “calibrating” and “balancing” the machine compensates for any ampoule defects or contamination, and reduces to minimal the chance of inac*487curate readings. The court concludes that the defendant has not demonstrated that preservation of the ampoules for re-examination purposes would offer a significant aid to establishing guilt or innocence, i.e., materiality.6
(3) Finally, there is no genuine dispute that preservation of breath samples by one of two available methods, the indium crimper (or breath encapsulator) and the silica gel trapping device, is not only possible, but is already in widespread usage throughout the United States.7 The underlying basis of breath preservation and retesting has met with wide scientific acceptance. (Frye v United States, supra; Garcia v District Ct., supra.) In addition, the entire procedure, from preservation through retesting, is simple, convenient, accurate and relatively inexpensive.8 The expert witnesses agreed that any of the problems with breath preservation were surmountable ones, just as there were potential problems in the operation of the otherwise accurate breathalyzer machine.
This narrows the focus from five issues to one, but it does not end the inquiry. If breath sample preservation is possible and inexpensive, and if it provides a means for the defendant to, in effect, “cross-examine” the most damaging witness against him, do the police have a duty to preserve such a sample for the defendant’s use?9
Both parties have submitted memoranda of law on the constitutional issues involved in preservation of breath *488samples, with particular emphasis on the landmark case of United States v Bryant (439 F2d 642).
UNITED STATES V BRYANT
The Bryant case grappled with a difficult problem in the criminal law — when evidence is destroyed, or simply not preserved by the police, the materiality or relevance of the evidence becomes in large part a matter of conjecture. Since the defendant cannot clearly demonstrate prejudice, why should the police be sanctioned for their malfeasance or nonfeasance? How would the police be deterred without sanctions, but what sanctions are appropriate under each set of circumstances?
The Supreme Court opinions such as Brady v Maryland (373 US 83), which involved suppression of existing evidence, were of no aid in this regard since the suppressed evidence was either demonstrably material or immaterial, crucial to guilt or innocence, or merely redundant. It has been pointed out10 that the problem with applying the principles of evidence suppression to cases of destroyed evidence is that the two are entirely different situations. The Bryant court attempted to resolve this dilemma by placing the burden on the government of demonstrating that earnest efforts, in the form of regular procedures had been made to preserve crucial evidence. “Earnest efforts” would be strictly construed against the People. Systematic nonpreservation of evidence might be regular, but it would be insufficiently protective of defendant’s due process rights. (United States v Bryant, 439 F2d 642, 652, n 21, supra.)
The logic of the Bryant decision was to focus on the procedure of preservation instead of materiality, since the latter could not be determined with precision. The underlying rationale is worthy of being quoted in full: “The purpose of the duty [of preservation] is not simply to correct an imbalance of advantage, whereby the prosecution may surprise the defense at trial with new evidence; rather, it is also to make of the trial a search for truth informed by all relevant material, much of which, because of imbalance in investigative resources, will be exclusively in the hands of *489the Government.” (United States v Bryant, supra, p 648; emphasis supplied.)
With the passing years, Bryant has attained the stature of a pronouncement of the Supreme Court, much as Frye has. Numerous jurisdictions have recognized in it a mandate for the preservation of breath samples. The only two reported New York cases have split on this issue. (Cf. People v Shepherd, 118 Misc 2d 365; People v Rich, 118 Misc 2d 1057.) As a general proposition, Bryant has been applied by the lower New York courts in a wide variety of situations, in the absence of a definitive ruling from the Court of Appeals.11
EVIDENCE IN POSSESSION
It has been argued by some, although not by the prosecutor in this case, that breath samples are not something to be preserved by the police, since they are never in the possession of the police. They are not now and have never been in existence. (See, e.g., People v Miller, 52 Cal App 3d 666.)
If one were to rely on semantics to determine this important question, one could argue that the police, in the operation of the breathalyzer, “seize” a breath sample from a suspect, and in passing it through an ampoule, “destroy” *490it, thus triggering the duty to “seize” an additional sample for the defendant’s use.
The court prefers to employ a more realistic approach. In this situation, the proper rule should be that the failure of the State to collect and preserve evidence, when those acts can be accomplished as a mere incident to a procedure routinely performed by State agents, is tantamount to suppression of that evidence. (Garcia v District Ct., 197 Col 38, 46, supra.) It is incumbent upon the State to employ regular procedures to preserve evidence which a State agent could reasonably foresee “might” be “favorable” to the accused. (Garcia v District Ct., supra; United States v Bryant, supra.)
THE SECOND CIRCUIT APPROACH
The People would eschew the Bryant approach, and have the court apply what they call the “pragmatic balancing test” of the Second Circuit Court of Appeals. In such cases as United States v Grammatikos (633 F2d 1013), the court weighed the degree of governmental negligence or bad faith involved, the importance of the evidence lost and the evidence of guilt adduced at the trial (quoting United States v Bryant, 439 F2d 642, 653, supra). The standard has also been described as weighing the extent of government culpability for failure to preserve against the amount of prejudice to the defendant. (United States v Sanchez, 603 F2d 381; United States v Bufalino, 576 F2d 446; see, especially, United States v Miranda, 526 F2d 1319,1327, n 5.)
However, in none of the cited Second Circuit cases was the destroyed evidence of the central importance and materiality of breath samples. In both Grammatikos (supra) and Miranda (supra), for example, the destroyed evidence was police “back up” tape recordings, as in Bryant. As even the Bryant court ultimately decided, this type of evidence was merely collateral to guilt or innocence; the destroyed tapes would have been of value in cross-examination, but, according to the court’s analysis, little else. In order to resolve the issue, the Second Circuit was obliged to engage in an excruciating examination of the possible materiality of the nonexistent evidence. The general result in these cases was a warning to the government to avoid such *491misconduct in the future, even while failing to apply sanctions in the present. (See, e.g., United States v Miranda, supra.) This is an appropriate disposition when the evidence is merely redundant and cumulative, where there was substantial additional evidence of guilt. (United States v Miranda, supra, pp 1328-1329.) Indeed, in a drug sale case, it is not necessary for the People to produce as part of their case a back-up tape recording; the testimony of the informant is enough. (See People v Palmer, 652 P2d 1092 [Col].)
By contrast, an alcoholic breath sample or blood sample is per se evidence of guilt. There is no substitute for a chemical test, which is a necessary part of the People’s case. Under subdivision 2 of section 1192 of the Vehicle and Traffic Law, the offense itself is having a specified percentage (.10%) of alcohol in the blood. Whether a driver is actually “intoxicated” in the sense of the old law, or of subdivision 3 of section 1192 is almost beside the point. (See People v Fox, 87 Misc 2d 210; People v Gmitter, 81 Misc 2d 745; People v LaMontagne, 91 Misc 2d 263; Taylor, Blood-Alcohol Analysis and the Fourth Amendment, 10 Search and Seizure L Rep, No. 3, p 121.)
When the evidence destroyed is almost literally the crime itself, the defendant’s need for independent testing and examination of the very gist of the crime become imperative.12 Thus, in People v White (40 NY2d 797) where the suppressed evidence was a controlled substance, it was reasoned, “Defendant’s guilt or innocence hung exclusively on the nature and amount of the substance in question; he advanced no other theory of defense. For refutation of the *492charge against him, there was no acceptable alternative to scientific testing by experts of his choice.” (40 NY2d, at p 798.) A legislative recognition of this right to independent testing is section 450.10 of the Penal Law, dealing with disposal of stolen property.
If the per se statute has strengthened the prosecution’s hand, so has the judicial approval of the breathalyzer machine. The foundation for admission of the breathalyzer need no longer include evidence of its inherent reliability, which is now an assumption. (See People v Gower, 42 NY2d 117, 121-122; People v Donaldson, 36 AD2d 37.)
The totality of circumstances has produced a fundamental injustice — the defendant may not confront the most damaging witness against him, except indirectly. He may challenge his blood alcohol reading only through the person who ran the machine; he may not challenge the machine itself. Yet, in most cases, the machine’s “testimony” will be dispositive.
The court is compelled to rule that basic fairness and the duty to make of the trial a search for truth informed by all relevant material require that the police preserve for the defendant’s use a separate breath sample for testing and analysis. There is no other substitute for this safeguard that adequately protects the defendant’s rights. While the intent of removing the intoxicated driver from our highways is laudable, we must not in our enthusiasm allow the due process rights of our citizens to be impaired. (See People v Farmer, 36 NY2d 386 [Fuchsberg, J., concurring opn]; see, also, People v Coutard, 115 Misc 2d 630 [overcharging by police].) The dangers involved are magnified when the underlying charge is a predicate to a felony.
DUE PROCESS
A denial of due process must be considered a failure to observe that fundamental fairness essential to the very concept of justice. (Kinsella v Singleton, 361 US 234.) Due process, unlike some legal rules, is not a technical conception with a fixed content unrelated to time, place and circumstances. (Cafeteria Workers v McElroy, 367 US 886, 895.) It is flexible and calls for such procedural protections as the particular situation demands. (Morrissey v Brewer, *493408 US 471, 481.) Clearly, advances in technology are one such change in circumstance. (See State v Cornelius, 122 NH 925, 929 [dissenting opn].)
The court is not persuaded by the People’s references to the good faith of the police officers in following their testing procedures, which do not comply with due process.13 The good faith of a government agent is no comfort to a defendant faced with no independent means of verifying the evidence against him; he is in as bad a situation as if the evidence had been maliciously fabricated. It is also obvious that good faith will often be difficult to determine; rarely will bad faith be admitted on the witness stand. (See Comment, Judicial Response to Governmental Loss or Destruction of Evidence, 39 U of Chicago L Rev 542.) To rely solely on the good faith of the police, who are engaged in the “competitive enterprise of ferreting out crime”, is to embark on a course that gradually will see the erosion of our due process freedoms.
For this part of the Criminal Court, this decision to suppress will be given full retroactive effect. “Where the major purpose of new constitutional doctrine is to overcome an aspect of the criminal trial that substantially impairs its truth finding function and so raises serious questions about the accuracy of guilty verdicts in past trials, the new rule has been given complete retroactive effect. Neither good-faith reliance by state or federal authorities on prior constitutional law or accepted practice, nor severe impact on the administration of justice has sufficed to require prospective application in these circumstances.” (Williams v United States, 401 US 646, 653; United States v Johnson, 457 US 537.)
The trial of a criminal charge should be a sober search for truth, not a sporting event. (Giles v Maryland, 386 US 66, 102; People v Copicotto, 50 NY2d 222, 226.) Many centuries ago, it was stated that in the search for truth, no man has yet been harmed.14 The prosecution of intoxicated drivers should be, at worst, only temporarily impaired by the duty to preserve breath. Even where the breathalyzer *494test results cannot be introduced, the charges of driving while impaired or intoxicated (Vehicle and Traffic Law, § 1192, subds 1, 3) can often be sustained on other evidence. It would be preferable for the good of society, and for the due process rights of defendants, however, that the New York City Police Department obtain one of the two types of breath-trapping devices displayed in court.15 Basic fairness requires that they do so. There can no longer be any question that driving while intoxicated is now a serious crime calling for serious safeguards. Should the police follow the lead of their brother officers throughout the country, the result will be only a temporary inconvenience for the administration of justice. A more permanent result will be the preservation of the integrity of the truth-finding process.
The motion to suppress the breathalyzer test results is granted.
. L 1981, ch 910, § 4.
. A nationwide crackdown on drunken drivers has produced a wide range of attacks on the breathalyzer and other blood alcohol-measuring machines:
State v Brockway (2 Ohio App 3d 227 [attack on reliability of intoxilyzer machine]); State v Lovato, (94 NM 780 [blood test sample destroyed by prosecution as deprival of due process]); Baca v Smith (124 Ariz 353 [duty of police to preserve a breath sample]); People v Hitch (12 Cal 3d 641 [chemical ampoules must be preserved for independent retesting and re-examination]).
. The Smith and Wesson model 900A breathalyzer machine uses glass ampoules containing a yellow-colored solution composed of potassium dichromate and sulphuric acid. When the alcoholic breath of the test subject is passed through this solution, the chemicals are bleached white in proportion to the intensity of the alcohol contained in the breath. A photoelectric cell in the breathalyzer machine compares the “test”, or bleached ampoule with a “control”, or unopened ampoule, and the machine assigns a numerical value to the degree of bleaching. This number is the blood alcohol reading. In the course of an examination, the operating technician examines the size of the ampoule by inserting it into a gauge, which also measures the volume of liquid contained. He also determines whether the machine is operating at the proper temperature for the chemical reaction to take place. Other test procedures include purging the machine’s air chambers with room air to prevent contamination, and a process known as “calibration”. Calibration is done by passing a solution of a known strength of .10 alcohol through the machine to determine if it is reading accurately. When a blood alcohol reading is derived, the technician merely reads it off a meter on the face of the machine and writes it down. Other models of the breathalyzer reportedly produce some sort of readout; the model 900A, in use throughout New York City since 1973, doesn’t.
. At the beginning of the hearing, counsel for defense withdrew a sixth ground based on radio frequency interference with the operation of the breathalyzer. (See 111 NJLJ 481; People v Hochheimer, 119 Misc 2d 344.)
. Dr. Cohen attempted to preserve breath samples in his refrigerator and freezer in accordance with the now fabled article by Jones and Volpe: Storage Properties of Breathalyzer Test Ampoules, Journal of Studies on Alcohol, vol 40, No. 11. (See People v Santiago, 116 Misc 2d 340 [Lang, J.].) Two of the five samples preserved by Dr. Cohen were bleached white, apparently by room air contamination, within a short period of time, and were rendered untestable. The three remaining ampoules deteriorated in a more predictable manner, a result which the expert for the People denigrated as merely “fortunate.” This court sees no reason, subject to the standards in Frye v United, States (293 F 1013) why Dr. Cohen’s experiment should be accorded the status of accepted scientific practice.
Even though it is true, as defendant’s counsel continually insisted, that every new idea must have its first day in court (United States v Franks, 511 F2d 25), the hearing demonstrated that there is still no general consensus on the scientific principles underlying preservation of ampoules for retesting. To a large degree, Dr. Cohen’s experiments confirm what has already been demonstrated about the unpredictable nature of the deterioration of ampoules. (See People v Santiago, supra; People v Le Free, 105 Misc 2d 1066; People v O’Brien, indictment No. 50107, County Ct, Nassau County, Jan. 29, 1981 [Santagata, J.]; see, generally, Ann., 19 ALR4th 509.)
. Preserved ampoules could conceivably be of some aid in cross-examining the police technician as to whether the proper steps were followed during the test, but this is not the same thing as materiality. (Cf. State v Canaday, 90 Wn 2d 808.) It is worth noting that while a preserved ampoule might demonstrate that there was defective good-faith testing by the police, it will not detect wholesale manufacture of evidence. By contrast, a preserved breath sample would be conclusive as to both problems. The court concludes that the benefits to be gained from ampoule preservation, taking into account the safeguards against inaccurate good-faith testing, are too speculative and slight to justify the burden of preserving thousands of such ampoules per year. The defendant has not met his burden of demonstrating materiality on these three issues.
. The following jurisdictions, by statute or case law, impose a duty upon the police to preserve breath samples: People v Trombetta (141 Cal App 3d 400); Municipality of Anchorage v Serrano (649 P2d 256 [Alaska]); Baca v Smith (124 Ariz 353); Garcia v District Ct. (197 Col 38); Vt Stat Ann, tit 23, § 1203, subd [a]; Okla Stat Ann, tit 47, § 752, subd 5.
. The price of a silica gel trap, or of an indium tube, was estimated at a few dollars. If the indium method was chosen, a crimping tool, priced at $300, would have to be purchased.
. Destruction or nonpreservation of evidence by the police is imputed to the People, since such nonpreservation taints the trial no less than any action of the prosecutor. (Barbee v Warden, 331 F2d 842; United States v Bryant, 439 F2d 642, 650.)
. Comment, Judicial Response to Governmental Loss or Destruction of Evidence, 39 U of Chicago L Rev 542.
. Lower New York courts have wrestled with the Bryant problem for many years without explicit guidance from the Court of Appeals. In cases where evidence has been merely suppressed, the State’s highest court has imposed a high duty of disclosure on the prosecution. “The duty arises out of considerations of elemental fairness to the defendant and as a matter of professional responsibility.” (People v Simmons, 36 NY2d 126, 131, quoting People v Savvides, 1 NY2d 554, 556.) In many cases, the suppressed evidence was deemed harmless, the evidence merely cumulative. (People v Codarre, 14 NY2d 370; People v Stridiron, 33 NY2d 287.) However, where the suppressed Grand Jury testimony might have altered the result at the suppression hearing, the court held that the prosecutor had denied the defendant his due process rights. (People v Geaslen, 54 NY2d 510.)
In People v Saddy (84 AD2d 175), the Second Department sanctioned the police for deliberate, though good-faith, erasure of a drug sale tape recording. There are other numerous instances where United States v Bryant (439 F2d 642) has been applied by the Appellate Division courts. (See, e.g., People v Bowdoin, 89 AD2d 986; People v Angelo, 93 AD2d 264; People v Watkins, 67 AD2d 741; but see People v Briggs, 81 AD2d 1017 [ruling that alcohol blood sample was not material to guilt or innocence].)
Lower court cases of interest include People v McCann (115 Misc 2d 1025), wherein the police failed to voucher the crime victim’s bloodstained pants. Citing Matter of Abe A. (56 NY2d 288), the court held that blood samples are material and exculpatory, and, accordingly, dismissed the indictment. Where the People failed to preserve the very essence of the crime charged, pursuant to their duty under United States v Bryant (439 F2d 642) and section 450.10 of the Penal Law, the only appropriate disposition of the charge was dismissal in People v Davis (109 Misc 2d 230). Examples abound, but these should suffice to demonstrate the continued efficacy of Bryant in New York.
. The People assert that the defendant has a right “to have a physician of his own choosing administer a chemical test” at the time of arrest to verify the breathalyzer results. (Vehicle and Traffic Law, § 1194, subd 8.) This right, it is argued, is adequate due process protection for any defendant victimized by faulty or bad-faith testing on the part of the breathalyzer operators.
The court disagrees. This “right” is, for all practical purposes, wholly illusory, even if one were to discount the fact that many of the defendants that come before this court lack any basic sort of health care, let alone a “physician of [their] own choosing”. First, the defendant would have to be aware of this right without being warned of it, since the police have no duty to do so. (See People v Hoats, 102 Misc 2d 386.) Secondly, the defendant would need the means to transport himself to a testing facility within the statutory two hours from arrest. The notion that a physician in this day and age could be prevailed upon to visit Highway 1 to conduct these tests hardly bears comment.
Where the practical difficulties concern evidence of this central importance, it can hardly be argued that the defendant’s due process rights are adequately protected by the statutory provision.
. There is no merit to the People’s remaining contentions, especially the notion that the court lacks the power to determine the constitutionality of police testing procedures.
. The quote is from Marcus Aurelius — Meditations VI, 21.
. It is suggested that a defendant could be given written notice at arraignment of his right, within 7 to 10 days after arraignment, to request independent testing of the breath sample. The testing would be done in the presence of defense counsel, under the auspices of the New York City Police Department. If the request was not made within the designated period, the sample could be destroyed, minimizing the storage burden on the police to the greatest extent possible.