OPINION OF THE COURT
Meyer, J.A violation of Vehicle and Traffic Law § 1192 (2) is not established unless the trier of fact finds that while operating a motor vehicle defendant had a blood alcohol content (BAG) of .10 of 1% or more. Evidence that a breathalyzer test administered within two hours of arrest showed defendant to have such a BAG is sufficient to establish prima facie a violation of the subdivision. It is, however, error not to permit defendant’s attorney to argue on the basis of evidence, whether through cross-examination of the People’s witnesses or testimony of defendant’s witnesses, expert or other, from which it could be found that defendant’s BAG at the time of vehicle operation was less than .10%, that if the jury so found defendant was not guilty of violating the subdivision. It is error also to charge that evidence that the individual administering the test possesses a permit from the Department of Health creates a rebuttable presumption that the breathalyzer examination was properly given, and to receive in evidence the logs for the breathalyzer, simulator and ampoule on the basis of foundation testimony which did not establish that it was the regular course of business to make the records contained in those logs at the time of the tests recorded in them or within a reasonable time thereafter (CPLR 4518 [a]). For all of those reasons, the order of the Appellate Term should be reversed and a new trial ordered.
I
Shortly before 1:28 a.m. on April 19, 1983, defendant’s automobile ran off the roadway of Bayville Avenue and partially through a fieldstone wall some 10 to 15 feet from the road. Officer Sprague, who responded to the radio run concerning the accident, found defendant behind the wheel, the windshield of the automobile cracked and defendant bleeding *140from lacerations of his face and nose. Asked what happened, defendant responded that he was not sure, that he remembered only driving along Bayville Avenue, but nothing after that. The officer noted that defendant had some trouble speaking and that as he exited the vehicle to enter the ambulance that had been called he staggered, that his eyes were bloodshot and that there was an odor of alcohol on his breath.
The officer, who had made a mental note to arrest defendant for driving while intoxicated, arrived between 1:55 and 2:00 a.m. at the hospital to which defendant had been taken. He could not administer physical coordination tests because defendant was being treated for his injuries, but he placed defendant under arrest and then called the Highway Patrol Bureau for a breathalyzer technician. Officer Needleman, responding to the call, arrived at the hospital at 2:37. Defendant was given the required warnings concerning submission to a breathalyzer examination and consented to do so, signing the required authorization form. Needleman also noticed that defendant was slurring words and that there was an odor of alcohol on his breath. Having determined that the breathalyzer was correctly set up and was functioning properly, Needleman had defendant blow into the instrument twice. The first sample, taken at 3:25 a.m., yielded a .15 reading. The second, taken at 3:35 a.m., after the instrument had been purged, recorded a reading of .16. On the basis of those readings defendant was charged with a violation of Vehicle and Traffic Law § 1192 (2). The maintenance, simulator and ampoule logs for the breathalyzer were admitted in evidence on the basis of Needleman’s testimony that they were made in the normal course of business and that it was the normal course of business to keep them, but over defendant’s objection that the People had to show "that the entries are made on [sic] or about the time that the events occurred”, which the Trial Judge overruled. Each log also bore the certification that it was a true copy of the report made by the breathalyzer technician of the Nassau County Police Department.1
Needleman testified that both tests were performed under identical circumstances and explained the difference between the two readings as occurring because not a deep enough *141sample was obtained at the two different times or because one sample was longer than the other. Through Dr. Jay Zimmerman, associate professor of physiology at St. John’s University, however, defendant presented evidence that alcohol is absorbed into the blood predominantly (90%) through the small intestine and that the time for absorption varies, depending on the presence of food, the type of alcohol, the type of mixer, the physical state of the individual and the presence of stress or trauma, from 15 to 40 minutes on the one hand to two to three hours on the other. Mixers and the presence of food, he explained, would delay the rate of absorption, fats being digested much more slowly, carbohydrates and proteins more rapidly. He further identified the food (a bacon cheeseburger with french fries), which defendant had testified he had eaten at about 12:30 A.M., as consisting of protein and carbohydrates but to a large extent fat, the presence of which "dramatically delays the absorption of alcohol”, and stated that it was possible that defendant’s blood alcohol would be on the increase two to three hours after such a meal and that there was no way to predict from the .15 and .16 readings testified to by Needleman whether defendant’s BAG was greater than .10 two hours earlier.
Notwithstanding Dr. Zimmerman’s testimony, when defendant’s attorney during summation began to argue that it was fair to infer from the fact that defendant’s BAG was rising at 3:35 that at the earlier time when he was driving it was much lower, the Trial Judge responded to the prosecutor’s objection by stating in the presence of the jury that the only question for the jury was whether the sample was taken within two hours of arrest and showed a BAG in excess of .10, that "[t]heir concern would be what was the reading at 3:25 and 3:35, that’s all.” His charge to the jury, however, was that to find defendant guilty of violating Vehicle and Traffic Law § 1192 (2) they had to find that "defendant operated a motor vehicle at the time and place in question and that at the time he had .10 of one percent or more by weight of alcohol in his blood” as shown by chemical analysis. And with respect to the analysis he charged that under Vehicle and Traffic Law § 1194 (9) the fact that it was made by a person possessing a permit issued by the Department of Health was presumptive evidence that the examination was properly given but that "presumptions may be rebutted or overcome by other evidence tending to disprove the presumption.” Defendant’s attorney excepted to both portions of the charge: as to the first because he had *142been prevented from so arguing in summation to the jury and because the charge would be misleading to the jury; as to the second because the statutory presumption was permissive in nature and the defense did not have the burden of rebutting it.
The jury found defendant guilty of violating Vehicle and Traffic Law § 1192 (2). On appeal by defendant from the judgment entered on that verdict, the Appellate Term affirmed in a memorandum which discussed only the scope of the subdivision. As to that it noted that "[t]he gravamen of the crime is the operation of a motor vehicle after ingestion of sufficient alcohol to produce the reading condemned by the statute within two hours of arrest”. The appeal is before us by leave of a Judge of this court (66 NY2d 765). We conclude (1) that defendant was improperly denied the opportunity of arguing to the jury that his BAG at the time of operation was less than .10, (2) that the breathalyzer logs were admitted without proper foundation and (3) that the charge with respect to Vehicle and Traffic Law § 1194 (9), because it may have been construed by the jury to impose upon defendant the burden of rebuttal, was improper. We therefore reverse.
II
Vehicle and Traffic Law § 1192 (2) provides that "[n]o person shall operate a motor vehicle while he has .10 of one per centum or more by weight of alcohol in his blood as shown by chemical analysis of his blood, breath, urine or saliva, made pursuant to the provisions of section eleven hundred ninety-four of this chapter.” The legislative history of the subdivision is traced in People v Schmidt (124 Misc 2d 102) and need not be repeated in detail here. It is sufficient for purposes of the present case to note that whereas prior versions of the section were phrased in terms of the prima facie effect of various BAG levels, the present subdivision, enacted by New York and, in varying versions, by some 28 other jurisdictions under congressional impetus,2 does not contain similar language. Some confusion has resulted from the fact that although section 1192 (2) proscribes operation while the driver has a BAG of .10 or more, Vehicle and Traffic Law § 1194 (1) permits the chemical test to be made "within two hours after such person *143has been placed under arrest for any such violation,” for, as Dr. Zimmerman’s testimony indicates and as other cases and commentators have noted, while blood alcohol content declines over time, the decline does not begin until sometime after the last drink, variously estimated as from 45 minutes to an hour and a half (Fuenning v Superior Ct., 139 Ariz 590, 680 P2d 121; State v Sutliff, 97 Idaho 523, 547 P2d 1128; People v Kappas, 120 Ill App 3d 123, 458 NE2d 140; State v Gallant, 108 NH 72, 227 A2d 597; Commonwealth v Hartman, 383 Pa 461, 119 A2d 211; 1 Erwin, Defense of Drunk Driving Cases § 15.03 [3d ed]; Brent & Stiller, Handling Drunk Driving Cases § 2.4; McCormick, Evidence § 205, at 615 [3d ed]; Fitzgerald & Hume, Single Chemical Test for Intoxication: A Challenge to Admissibility, 66 Mass L Rev 23; Hume & Fitzgerald, Chemical Tests For Intoxication: What Do The Numbers Really Mean?, 57 Analytical Chemistry 876A). A test reading taken within two hours after arrest may, therefore, be higher or lower than the actual BAG while driving.
Predicated on that fact, defendant argued before the trial court that it was the State’s obligation in order to make a prima facie case to relate the .10 BAG back to the time of operation.3 Before us, however, he contends only that the Trial Judge erred in refusing to allow him to argue to the jury that in fact defendant’s BAG at the time of operation was less than .10. The cases in which the question has been considered so far in New York have not been consistent in their holdings. People v Ippolito (100 AD2d 734) approved defendant’s being permitted to "present expert testimony that the breathalyzer test was not accurate”, both Ippolito and People v Molina (128 Misc 2d 638 [App Term]) recognized that breathalyzer results constitute strong but not conclusive evidence of guilt, and the holding in People v Zigles (116 AD2d 610, lv denied 67 NY2d 767) implies as much. On the other hand, the Appellate Term in the present case and in People v LaMontagne (91 Misc 2d 263), reading the subdivision expansively, as though it prohibited operation not while the driver has a .10 BAG, but when the driver has ingested sufficient alcohol to produce a .10 BAG within two hours thereafter, has ignored the scientifically *144accepted, fact that a .10 reading within two hours after operation does not establish a .10 reading while operating and, in effect, has rewritten the statute. And the Criminal Court of the City of New York in People v Darbouze (122 Misc 2d 654) held a BAG reading of .10 "per se evidence of guilt” (122 Misc 2d, at p 655) and in People v Schmidt (supra), characterized the subdivision as "a .10 per se liability standard” (124 Misc 2d, at p 108) and a ".10 absolute liability provision” (id., at p 112), pointing out that charts4 are readily available which indicate, with a reasonable degree of certainty, the number of alcoholic beverages necessary for a particular individual to reach a BAG of .10 (id., at p 113), but somewhat inconsistently held that "what is required is that the People prove, by objective, scientific criteria that at the time defendant was driving, his BAG was .10%” (id., at p 108).
When one turns to out-of-State cases, care must be taken to distinguish the language of the relevant statutes. Thus Burg v Municipal Ct. (35 Cal 3d 257, 673 P2d 732, cert denied and appeal dismissed 466 US 967) and Fuenning v Superior Ct. (139 Ariz 590, 680 P2d 121, supra) both involved statutes simply making it unlawful to drive while there is .10% or more by weight of alcohol in the driver’s blood. Erickson v Municipality of Anchorage (662 P2d 963 [Alaska]); Lovell v State (283 Ark 425, 678 SW2d 318); Coxe v State (281 A2d 606 [Del]); and State v Franco (96 Wn 2d 816, 639 P2d 1320) concerned statutes, like Vehicle and Traffic Law § 1192 (2), making it unlawful to drive with a .10 BAG "as shown by” or "as determined by” a chemical test within a specified time, while State v Tanner (15 Ohio St 3d 1, 472 NE2d 689) and State v Ulrich (17 Ohio App 3d 182, 478 NE2d 812) dealt with a statute the first subdivision of which was phrased like the California and Arizona statutes referred to above, but the second subdivision of which provided that "the court may admit evidence [as to the BAG] at the time of the alleged violation as shown by chemical analysis of the defendant’s * * * breath * * * withdrawn within two hours of the time of such alleged violation” (emphasis supplied).
Both Burg and Fuenning recognized the per se nature of the offense but held that conviction required the jury to find that defendant’s BAG at the time of driving was .10% or more and *145recognized defendant’s right to challenge whether the test result accurately measured defendant’s BAG at the time of operation. Coxe v State (supra) interpreted the "as shown by” type of statute to require the State to prove "only that the defendant was in physical control of the vehicle, and that a proper and timely test showed the required percentage of alcohol concentrated in the defendant’s system” (281 A2d, at p 607), but did not consider whether the test result could be rebutted by defendant, and State v Ulrich (supra) held the evidentiary subdivision authorizing admission of the test result to be a legislative determination relating the test result back to the time of driving but likewise had no reason to consider whether it was conclusive.
Lovell (accord, Long v State, 284 Ark 21, 680 SW2d 686) held it a per se violation to drive with a BAG of .10 but that its "as determined by” language did not lessen the State’s burden of proof that defendant was driving with a .10 BAG. Erickson held that the breathalyzer result alone established a prima facie case as to defendant’s BAG while she was driving but construed the "as determined by” language of the Alaskan statute to be intended to establish a presumption that a breathalyzer test conducted within four hours of driving will result in a reading equal to or less than the actual BAG while driving and noted that even if the accused offered no rebutting evidence the court should instruct the jury in terms of permissive inferences that it may but is not bound to infer defendant’s BAG while driving from his or her BAG at test.
The Ohio courts, in Tanner and Ulrich, held that their statute did not create a presumption but rather defined the conduct prohibited, but held that defendant could challenge his specific test result and that the jury should consider not only that result but all other relevant evidence in determining whether the State had shown beyond a reasonable doubt that defendant had committed a violation. And in Franco, the Washington Supreme Court, agreeing that its statute did not establish a presumption, answered defendant’s contention that by being required to give a blood sample he was being forced to testify against himself, by holding, in reliance on Schmerher v California (384 US 757) and similar cases, that giving a blood sample is not testimonial in nature. It then went on to state, however, that: "The defense has the same opportunity to attack that reading as they always have had under the prior presumptions. The defense is entitled to an expert witness instruction which was offered in this case by Franco *146and refused by the court, in error. Additional expert testimony, while available to the defendant, is not the only method of impeaching the reading on the Breathalyzer. The State’s expert testimony may be controverted by the defendant testifying about the number of drinks he consumed and the effects of the alcohol upon him, he may call lay witnesses to testify as to those same factors, he may argue that the machine must be in error because of the slight effect the alcohol had upon him. It is simply not the case that the giving of the breath sample proves the crime.” (96 Wn 2d, at pp 828-829, 639 P2d, at pp 1326-1327, nn omitted.)
Clearly the above cases cannot all be reconciled but they suggest the difficulty arising from the absolutist construction given the statute by the Appellate Term. In our view it must be construed bearing in mind that statutes are to be construed so as to avoid doubt as to their constitutionality (McKinney’s Cons Laws of NY, Book 1, Statutes § 150 [c]). What the statute proscribes is operation of a motor vehicle "while [the operator] has .10 of one per centum or more by weight of alcohol in his blood”. As Dr. Zimmerman’s testimony established and as held in People v Ippolito and People v Molina (supra), and the out-of-State cases cited above, the BAG count shown within two hours after arrest is strong but not conclusive evidence of the BAG during operation. To foreclose a defendant’s introduction of evidence seeking to establish that his BAG while operating was less than .10 may raise doubt as to constitutionality (see, Ulster County Ct. v Allen, 442 US 140) and, as already noted, to construe it as has the Appellate Term is essentially to rewrite the statute. We conclude, therefore, that proof of a breathalyzer reading of .10 or more within two hours after arrest establishes prima facie a violation of Vehicle and Traffic Law § 1192 (2) which, together with evidence of one or more of defendant’s deportment, speech, stability and the odor of his or her breath, is sufficient to sustain a conviction, absent evidence, expert or other and by whichever party produced, from which the trier of fact could conclude that defendant’s BAG at the time of vehicle operation was less than .10. When, however, as here, such evidence has been presented, defendant must be permitted to argue its significance to the jury. Because he was foreclosed from doing so and because the court’s ruling during defendant’s attorney’s sum-*147motion and its instructions at the close of the case were in conflict on this issue, there must be a reversal.5
Ill
Admission of the breathalyzer logs over objection that it had not been shown that the entries were made at the time of the acts recorded in them or within a reasonable time thereafter was also error. CPLR 4518 (a) expressly requires such foundation evidence. People v Farrell (58 NY2d 637) cannot be deemed to have sanctioned admission of records not shown to have been made at or within a reasonable time after the event, for the memorandum decision makes no mention of the point, which apparently was not argued. The statement in the Farrell memorandum that (at p 639) "the records list the date the analyses were performed” concerned the completeness of the documents as business records, not that the entries were contemporaneously made; the date of analysis would be stated even if the entry had been made long after the event it recorded.
Nor can we accept the People’s argument that the certificates were nevertheless properly admitted because each bore the certification set forth in footnote 1 above. As was held in Liberto v Worcester Mut. Ins. Co. (87 AD2d 477, appeal dismissed 57 NY2d 955, lv denied 58 NY2d 824), and as is made clear by the Eighth Annual Report of the Judicial Conference to the Legislature on the CPLR (1970),6 on the basis of which subdivision (c) was added to CPLR 4518, admissibility under the latter subdivision is governed by the same standards as the general business record exception in subdivision (a). Thus, *148a certificate made under CPLR 4518 (c) which does not set forth that the entries in the certified record were made at the time of the events they record or within a reasonable time thereafter is not admissible under that subdivision.
The People argue also that the certificates together with Officer Needleman’s testimony that the instrument was functioning properly was sufficient foundation for admission of the test results. We disagree for the reasons set forth in People v Freeland (68 NY2d 699 [decided herewith]) and because, as Officer Needleman himself testified, testing is done by certified mechanics while a breathalyzer technician such as Needleman is trained how to operate the instrument, not in testing and repairing it. While the scientific reliability of breathalyzers in general is no longer open to question (People v Gower; 42 NY2d 117; People v Donaldson, 36 AD2d 37), there must still be either proper foundation testimony under CPLR 4518 (a) or a proper CPLR 4518 (c) certificate to establish that the particular instrument used to test a defendant’s BAG and the ampoules used with it had been tested within a reasonable period in relation to defendant’s test and found to be properly calibrated and in working order (People v Freeland, supra; People v Todd, 38 NY2d 755).
IV
Vehicle and Traffic Law § 1194 (9) provides that proof that the chemical analysis was made by an individual possessing a permit issued by the Department of Health is "presumptive evidence” that the examination was properly given. Presumptive evidence, is, however, like the prima facie evidence to which CPLR 4518 (c) refers, evidence which permits but does not require the trier of fact to find in accordance with the "presumed” fact, even though no contradictory evidence has been presented (Matter of Commissioner of Social Servs. [Patricia A.] v Philip De G., 59 NY2d 137, 140). It is, in short, not a presumption which must be rebutted but rather an inference, like the inference of negligence denominated res ipso loquitur (see, George Foltis, Inc. v City of New York, 287 NY 108; Richardson, Evidence § 56 [Prince 10th ed]).
Because the jury may have understood from the charge as given that they were required to accept Needleman’s tests as properly given in the absence of testimony that they were not, the effect of the instruction was to shift the burden of proof to defendant (Ulster County Ct. v Allen, 442 US 140, supra; see, *149Sandstrom v Montana, 442 US 510; People v Guilford, 20 AD2d 192). It was, therefore, reversible error.
Accordingly, the order of the Appellate Term should be reversed and a new trial ordered.
. The certificates each read: "In accordance with Section 4518 of the C.P.L.R, I hereby certify that I am a designated clerk with the Nassau County Police Department and that this is a true copy of a report made by a breathalyzer technician of the Nassau County Police Department. Dated: October 27,1983.”
. 23 USC § 408 (e) (1) (C) makes enactment of a .10% BAG law mandatory for States wishing to obtain Federal funds to support their alcohol traffic safety programs.
. Only one State has so held (State v Rollins, 141 Vt 105, 444 AD2d 884; see also, State v Dumont, — Vt —, 499 A2d 787). Dictum to like effect will be found in People v LaPlante (81 Misc 2d 34) and People v Jones (77 Misc 2d 33). Under an early version of the statute it was held that "intoxication and operation must be simultaneous or there is no crime” (People v Strauss, 260 App Div 880, 881).
. Such a chart is reproduced in State v Tanner (15 Ohio St 3d 1, 4, 472 NE2d 689, 692; but see, Hume & Fitzgerald, Chemical Tests For Intoxication: What Do The Numbers Really Mean?, 57 Analytical Chemistry, at 878A).
. The People argue that defendant’s evidence was before the jury and was rejected by it, but it is simply impossible to know whether they found defendant guilty on the basis of the Judge’s comment during summation that all that counted was the BAG at 3:25 and 3:35 a.m. or because they concluded contrary to Dr. Zimmerman’s testimony that defendant’s BAG at 1:28 a.m. was .10 or more.
. "It is, of course, intended that the admissibility under this subdivision of the contents of the pertinent records would be governed by the same standards of relevancy which govern admissibility under subdivision (a).” The Eighth Annual Report is reprinted in the Sixteenth Annual Report of the Judicial Conference (1971), at page A-42, and in 1970 McKinney’s Session Laws of NY, at 2795-2796. (See also, 5 Weinstein-Korn-Miller, NY Civ Prac f 4518.26.) No argument has been presented to us as to possible admissibility of these records under CPLR 4520 (compare, People v Hoats, 102 Misc 2d 1004, with People v Drumm, 122 Misc 2d 1051), and we, therefore, do not pass upon that question.