OPINION OF THE COURT
William Leibovitz, J.The defendant challenges the validity of his breathalyzer reading taken less than one hour after his arrest on a charge of driving while intoxicated (Vehicle and Traffic Law § 1192 [2]), an unclassified misdemeanor. He advances two grounds for suppression or preclusion of the reading, which registered his blood alcohol content (BAG) at .18%.
*735First, the defendant disputes the accuracy of the so-called conversion ratio by which the breathalyzer machine transposes one’s breath alcohol measurement into blood alcohol content. Second, he attacks the reliability of the test ampoules and simulator solution used in the breathalyzer machine. Essentially, defendant’s argument in both instances is that under present circumstances the potential for error in breathalyzer testing is sufficient to exclude his BAG reading from evidence.
ACCURACY OF CONVERSION RATIO
The defendant asserts that his breathalyzer test result should be suppressed on the ground that the breathalyzer machine is based on a faulty premise and is therefore unreliable. The breathalyzer functions on the presumption that alcohol in the breath and blood are proportional to each other by a fixed ratio.
The breathalyzer is programmed to convert breath to blood alcohol content by applying a uniform ratio or 2,100:1 on the assumption that at normal body temperature the amount of alcohol in 2,100 milliliters of breath equals the alcohol in 1 milliliter of blood, referred to as the partition ratio (Gerstenzang, Handling the DWI Case in New York § 29.3, at 181 [1989 Supp]; 2 Erwin, Defense of Drunk Driving Cases § 18.01, at 18-5, 18-6 [3d ed 1989]).
The defendant argues that his breathalyzer reading should be suppressed because the fixed ratio of 2,100:1 is used by the breathalyzer to convert breath alcohol to blood alcohol, whereas in fact individual ratios differ. As a result, the fixed ratio overstates the blood alcohol level of a significant portion of the population. As an alternative to suppression, the defendant contends that this reading should be reduced by 52.38%, which he alleges to be the margin of error of the conversion ratio. The defendant does not present facts as to his individual ratio but relies on evidence in general that such variations do occur.
The People reply that the fixed 2,100:1 ratio has been proven to understate rather than overstate the average ratio of the general population, meaning that breathalyzer readings on the average show a lower blood alcohol content than is actually present. Absent any facts of this defendant’s individual ratio, and allowing for any recognized margin of error, the People state that the defendant’s .18% BAG reading is reliable *736evidence for the trier of fact to consider in determining that the defendant was driving while intoxicated as charged.
At the time the breathalyzer was introduced, it was believed that 2,100:1 was the constant breath/blood ratio of all persons (Gerstenzang, op. cit., at 181; Erwin, op. cit., at 18-5 — 18-7). Leading researchers have since learned that in fact partition ratios vary among individuals, that a single individual’s ratio is also subject to change, and that the mean population ratio is about 2,300:1. Factors which may alter one’s ratio include body temperature, hematocrit (red blood cell concentration), state of alcohol absorption, and other conditions (2 Dubowski, Recent Developments in Alcohol Analysis [1986]).
According to one expert witness, the ratio varies among individuals from 1,100:1 to 3,400:1 (State v Burling, 224 Neb 725, 400 NW2d 872, 876 [1987]). Another researcher has found ratios varying from 1,100:1 to 3,200:1 (State v McGinley, 229 NJ Super 191, 550 A2d 1305, 1311 [Sup Ct, Law Div 1988], revd sub nom. State v Downie, 229 NJ 207, 550 A2d 1313 [App Div 1988]). Dr. Kurt W. Dubowski, a noted authority, places ratio differences between 1,555:1 and 3,005:1 (Erwin, op. cit., at 18-19). He concludes that the breathalyzer with its fixed ratio of 2,100:1 understates the blood alcohol of 86% of the population and overstates it for 14% of those tested op. cit., at 18-64). Other experts assert that for 95% of the population the breathalyzer ratio of 2,100:1 results in a blood alcohol reading that is either accurate or understated (People v Pritchard, 162 Cal App 3d Supp 13, 16, 209 Cal Rptr 314, 316 [1984]).
Courts outside of New York have recently adopted a variety of responses to breath/blood alcohol testing. In State v Burling (supra, at 877), the Nebraska Supreme Court reasoned that because of the difference between their intoxilyzer ratio of 2,100:1 and the lowest stated partition ratio of 1,100:1, the defendant’s breathalyzer reading must be reduced by that difference of 52.38%. In cases subsequent to Burling, the court has held that the trier of fact must judge each case separately to determine whether credible evidence requires an adjustment, and its degree (People v Babcock, 227 Neb 649, 419 NW2d 527 [1988]). For that purpose, general evidence is admissible, and variable partition ratios of the population, as well as other general evidence, may be proven to show possible error in the fixed ratio of 2,100:1 used by the test machine (State v Hvistendahl, 225 Neb 315, 405 NW2d 273, 275-276 [1987]).
*737Until recently California followed a "rule of convenience,” similar to an affirmative defense, which requires the trier of fact to presume a defendant to have a 2,100:1 partition ratio unless the defendant presents evidence as to his personal ratio which establishes that the 2,100:1 ratio is not valid for him. General evidence of the possibility of error in the 2,100:1 ratio is insufficient to rebut the presumption (People v Pritchard, supra; People v Gineris, 162 Cal App 3d Supp 18, 209 Cal Rptr 317 [1984]; People v Herst, 197 Cal App 3d Supp 1, 243 Cal Rptr 83 [1987]).
However, in December 1988, a California court rejected the rule of necessity and held that general evidence should have been fully considered by the jury in the form of defense expert testimony that a person’s partition ratio fluctuates and is not constant, from which the jury could determine whether such general evidence raises a reasonable doubt as to the presumed 2,100:1 ratio (People v McDonald, 206 Cal App 3d 877, 254 Cal Rptr 384 [4th Dist 1988]).
In an intoxicated driving case still pending in New Jersey, an intermediary appellate court was reversed on its ruling that the trial court must consider expert testimony as to disparities between the fixed breathalyzer ratio and ratio variations of individuals generally (State v Downie, 229 NJ 207, 550 A2d 1313 [App Div 1988], supra, revg State v McGinley, 229 NJ Super 191, 550 A2d 1305 [Law Div 1988], supra). Reversal was based solely on a New Jersey Supreme Court order that in all cases judicial notice must be taken of the scientific reliability of the breathalyzer (Romano v Kimmelman, 96 NJ 66, 474 A2d 1 [1984]). In these cases no distinction was made between the admissibility of evidence showing variations in ratios generally and proof of an individual defendant’s variation from the breathalyzer ratio.
None of the foregoing State courts have considered or treated the conversion ratio issue in terms of suppression of evidence. In all cases, the breathalyzer test results were admitted and the issue presented was the weight to be given to such evidence by the trier of fact.
This court likewise concludes that the defendant does not allege facts that, if true, are sufficient to support a ground for suppression of the breathalyzer reading based on the claimed inaccuracy of the conversion ratio, and the motion to suppress is denied (CPL 710.60 [3]). For the same reason, the motion to reduce the breathalyzer reading is denied. Instead, the ques*738tion of the conversion ratio is one for the trier of fact to consider in evaluating the over-all weight to be given to the breathalyzer result.
The facts presented by the defendant as to variations in the partition ratio do not demonstrate that the breathalyzer conversion ratio of 2,100:1 requires either suppression or reduction of breathalyzer readings as a matter of law. The consensus of leading researchers is that 86%-95% of the entire population are reliably within that ratio. Any evidence in general or evidence as to a defendant’s individual ratio that would challenge the accuracy of the conversion ratio in a particular case can be tested by the fact finder in the crucible of a trial. Although the reliability in general of the breathalyzer is no longer open to question, its accuracy in any particular case is always open to investigation at the trial (People v Alvarez, 70 NY2d 375 [1987]; People v Mertz, 68 NY2d 136 [1986]).
The People must prove beyond a reasonable doubt at trial that the breathalyzer was properly administered. As a matter of fundamental fairness the defendant may explore any malfunctions, such as faulty calibration (People v Todd, 38 NY2d 755 [1975]; California v Trombetta, 467 US 479 [1984]) or other evidence indicating that the breathalyzer test reading was not accurate (see, People v Mertz, supra). As to any trial evidence challenging the accuracy of the calibrated fixed conversion ratio, the People must overcome such evidence beyond a reasonable doubt.
The trier of fact should consider any relevant general evidence concerning the ratio, including varying ratios of the population, to determine reasonable doubt. Any specific evidence of the defendant’s personal ratio would of course be relevant, but general evidence should not be precluded in the absence of such specific evidence.
This court believes that the courts of New Jersey and California, which have ruled that general evidence is insufficient to rebut the "presumption” against error in the 2,100:1 ratio, as previously discussed, have incorrectly removed that question from the trier of fact (Ulster County Ct. v Allen, 442 US 140 [1979]). Like all matters of science, the breathalyzer is subject to miscalculations, therefore requiring openminded inquiry. In the 2nd century A.D., Ptolemy confidently announced that the sun revolves around the earth, which remained unquestioned by science until 1543 when Copernicus *739proved that in fact the earth revolves around the sun. In 1952, seven leading experts agreed in a joint statement that: "The Alcometer, the Intoximeter and Drunkometer * * * will give comparable and reliable results for estimating the concentration of alcohol in the blood.” In 1972, several of the original signers adopted a new statement recognizing that all three named devices had been proved to be inaccurate and removed from use (Erwin, op. cit., at 18-6, 18-7).
Because the breathalyzer relies on principles of alleged scientific constancy, all evidence indicating variations or exceptions to these principles should be scrutinized by the fact finder. The People are free to disprove, through their experts, all general or specific evidence which the defendant contends are possible sources of error.
The defendant’s motion to suppress or reduce the breathalyzer reading based on the alleged inaccuracy of the conversion ratio is denied.
RELIABILITY OF BREATHALYZER CHEMICALS
The defendant claims that substances used in his breathalyzer test are unreliable and that his reading should be suppressed. In support he cites the December 1987 investigative report of the Pennsylvania Auditor General (the Report), which concluded that Systems Innovation, Inc. (SSI), the contract manufacturer of chemical components for breathalyzers used in New York State, has faulty laboratory procedures that "do not meet national scientifically acceptable quality control standards” (Report, at 3).
Especially criticized by the Report are SSI’s methods for manufacturing two products essential to accurate breathalyzer testing. The first is simulator solution, an alcohol and water mixture used to calibrate and verify the accuracy of the breathalyzer. The second product is the ampoule, two of which are used for each breath test. Ampoules are glass tubes containing three chemical compounds in distilled water, all of which must be mixed in exact premeasured proportions to achieve a correct breathalyzer result.
In this case the People have confirmed that SSI supplied the ampoules and simulator solution used in defendant’s breathalyzer test.
The method accepted by New York State for safeguarding the integrity of both products is that SSI assigns a lot number to the product as it is prepared in bulk. A sample of that lot is *740sent by SSI to the New York State Police Crime Laboratory for approval. If approved, the bulk is used by SSI to supply simulator solution and ampoules, bearing the applicable lot number, for breathalyzer testing throughout New York State. For each breathalyzer test, the police lab issues certificates of analysis for both products, confirming that a sample of that lot was chemically analyzed and approved. The People then seek to introduce these certificates of analysis to establish that the chemicals used in the ampoules and simulator solution were of the proper kind and mixed in the proper proportion in order to lay the necessary foundation for the admission of the test results.
According to the Report, investigation disclosed that production procedures at SSI are haphazard, record keeping is poor, and quality control is so inadequate as to suggest that lots are improperly interchanged, making the testing of samples invalid. In addition, the Report finds a lack of records at SSI to confirm the necessary identification, quantity and calculations as to the chemical contents of the products used for breathalyzer testing.
The People deny these claims of unreliability and refer to the fact that on November 2, 1988 the Pennsylvania Attorney General, after investigation, repudiated the Report in a letter to the Auditor General, which stated: "We have concluded that there is no reason to believe that the solutions provided by SSI were defective.” In addition, the People have presented the affidavits of three persons from police laboratories in New York which tested ampoules and simulator solution from SSI’s bulk deliveries in the field, including lots used in this case, and found them to meet the required standards of quality. Finally, the People argue that various alleged defects in the processes used by SSI are either provably nonexistent or scientifically inconsequential.
A breathalyzer reading is admissible at trial on this charge only if the People present "evidence from which the trier of fact could reasonably conclude” that the machine was working properly when the test was given and "that the chemicals used in conducting the test were of the proper kind and mixed in the proper proportions” (People v Freeland, 68 NY2d 699, 700 [1986]; People v Donaldson, 36 AD2d 37, 41 [4th Dept 1971]). These are the same issues raised by the defendant’s factual claims and the People’s response in the present case.
The People untenably assert that the defense has not suffi*741ciently alleged particularized facts to warrant a pretrial suppression hearing. Without suggesting that defendant’s claims and those contained in the Report will be sustained, the court nevertheless concludes that defendant has made specific factual allegations which, if true, support defendant’s grounds for suppression of the breathalyzer reading, therefore requiring a pretrial evidentiary hearing (CPL 710.60[4]; People v Patterson, 129 AD2d 527, 528 [1st Dept 1987]; People v Werner, 55 AD2d 317, 319 [4th Dept 1977]; cf., People v Carrion, 68 AD2d 827 [1st Dept 1979]).
Based essentially on investigation described in the Report, defendant alleges that SSI supplied the chemical materials identified by specific lot numbers for his breathalyzer test, but SSI kept insufficient records to show that laboratory-approved samples were from the same lots as the bulk product used in his test. Likewise, he alleges that the critical contents and proportions of chemicals in the ampoules for his test cannot be documented and are unreliable because of faulty processes and record keeping by SSL These are sufficiently particularized allegations in support of defendant’s grounds for suppression and cannot properly be likened to the general allegations in People v Garneau (120 AD2d 112 [4th Dept 1986]), cited by the People, where the defendant merely speculated that radio frequency interference could possibly affect a breathalyzer result, without claiming any specific negative effect in his case.
The question of reliability in this case is fundamental to whether the breathalyzer reading should be admissible in evidence at the trial. In other words, unless the People can establish that the certificate of analysis for the ampoules accurately reflects the chemical composition of the ampoules used to test the defendant, the breathalyzer test result will not be admissible. This is distinguishable from a trial issue in which the jury itself determines the weight to be accorded to evidence. A pretrial hearing is appropriate to evaluate admissibility and to avoid the all but impossible task of requiring the jury first to consider and then to ignore the evidence if found to be inherently unreliable.
At this hearing the People must initially come forward and establish that the breathalyzer was properly operated in obtaining defendant’s reading. On this issue the defendant does not challenge the reliability of the breathalyzer device in general, now universally accepted, and the People need not produce expert testimony as to the scientific principles under*742lying the breathalyzer (People v Mertz, supra, at 148; People v Donaldson, supra, at 40). The burden of proof will then shift to the defendant to prove his grounds for suppression by a preponderance of the evidence.
Although similar in result, this burden of proof is not derived from the standard imposed in suppression motions brought under CPL 710.20 relating to evidence unlawfully seized (see, e.g., People v Berrios, 28 NY2d 361, 367 [1971]). In the present case the burden of proof stems more logically from judical experience in determining the scientific reliability of various techological devices and the admissibility of such evidence.
Since Frye v United States (293 F 1013 [DC Ct App 1923] [admissibility of polygraph]), the test of admissibility of technological evidence is whether that technique is generally accepted in the scientific community (United States v Alexander, 526 F2d 161, 163 [8th Cir 1975] [polygraph]; United States v Stifel, 433 F2d 431 [6th Cir 1970] [neutron activation analysis]).
At a Frye hearing the party offering evidence of a technological device has the burden of proof. However, the breathalyzer and related devices have, in effect, already met the Frye test of general acceptance in the scientific community (People v Mertz, supra; California v Trombetta, supra, at 485 [intoxilyzer]). In the present case, the normal Frye hearing will, in a sense, be inverted. Here the defendant rightfully has the burden of proof because he alleges that a generally accepted breathalyzer device, as used in his particular case, produced a scientifically unreliable result.
The defendant’s burden of proof is not inconsistent with People v Hughes (59 NY2d 523, 547 [1983]), which requires the People to prove at a pretrial suppression hearing by "clear and convincing” evidence that hypnosis has not caused the prehypnotic recollections of a prosecution witness to be unreliable and inadmissible. The People’s burden in Hughes flows from the fact that hypnosis is not at present accepted under Frye standards. Conversely, the defendant’s burden in the instant case comports with the scientific community’s general acceptance of the breathalyzer.
This branch of defendant’s motion to suppress is granted only to the extent that a pretrial hearing is ordered.