Opinion

Defendant was convicted by a jury of violating Vehicle Code section 23152, subdivisions (a) (driving under the influence of alcohol) and (b) (driving with a blood-alcohol level of 0.10 percent or more). He appeals from the judgment, asserting errors in instruction. We find instructional error, and therefore reverse and remand for retrial.

Definition of “Under the Influence”

The jury was instructed that a person is under the influence of an alcoholic beverage when as a result of drinking such alcoholic beverage, his physical or mental abilities are impaired to such a degree that he no longer has the ability to drive a vehicle with the caution characteristic of a sober person of ordinary prudence, under the same or similar circumstances. Defendant requested that the court further instruct: “The word ‘sober’ means moderate in, or abstinent from, the use of intoxicating liquor. Synonyms for ‘sober’ are ‘abstemious’, ‘abstinent’, ‘temperate’, and ‘unintoxicated’ [szc].” The court refused.

Defendant argues that the jury could have been misled so that they measured impairment by comparison with a person who has not consumed any alcoholic beverage, rather than one who has consumed alcohol in moderation. But a person who has not consumed alcohol and therefore is not impaired by alcohol is the proper standard for comparison.

If defendant means the comparison should be to a person who has consumed alcohol but is not impaired, he makes a distinction without a difference, i.e., the standard for comparison—unimpaired by alcohol—is the same as a person who is unimpaired because he has not consumed any alcohol.

If defendant means the comparison should be to a person who has consumed alcohol and is impaired, but not as much as if he were under the influence, the instruction would be circular and erroneous. To define B as more than A, while defining A as less than B, does nothing to define either and leaves no standard for comparison.

*Supp. 16 The instruction makes it clear that the test is not whether there has been any consumption of alcohol, but whether there has been any impairment as a result.

The instruction, without further definition of a “sober person,” was approved in People v. Schoonover (1970) 5 Cal.App.3d 101 [ 85 Cal.Rptr. 69 ]. The trial court properly rejected defendant’s proffered instruction.

Burden Regarding Partition Ratio

Defendant submitted to a breath test on an Intoxilyzer, to determine the alcohol in his blood. Breath test machines convert the percentage of alcohol in the breath to alcohol in the blood by using a breath-blood “partition ratio” of 2,100 to 1, as required by the California Code of Regulations.

In November 1987, People v. Herst (1987) 197 Cal.App.3d Supp. 1 [ 243 Cal.Rptr. 83 ] was decided. The court approved an instruction which was in part as follows: “The jury may infer that this defendant has a 2100 to 1 partition ratio unless she presents evidence as to her personal partition ratio which establishes that the 2100 to 1 ratio does not apply to her. General evidence of such a possibility of error will not suffice to rebut this presumption. The People are not required to prove this Defendant’s particular partition ratio. [j[] The jury should not consider any conclusions made by any witness regarding the Defendant’s blood alcohol concentration based upon breath alcohol results which use a partition ratio other than 2100 to 1.” (Id., at p. Supp. 4, fn. 1.)

This case was tried in June 1988. The trial court instructed the jury: “It is not necessary for the People to prove that the defendant did not [s/c] have a partition ratio in the normal range. Whether the defendant has a particular partition ratio is peculiarly within his own knowledge. The burden is on the defendant to raise a reasonable doubt in that regard.”

In December 1988, People v. McDonald (1988) 206 Cal.App.3d 877 [ 254 Cal.Rptr. 384 ] was decided. In view of the evidence in McDonald , the court held that “it was error to instruct the jury that it should presume [defendant] has a 2,100:1 partition ratio unless he presented evidence as to his personal ratio which either proved the contrary or raised a reasonable doubt.” (Id. at p. 884.) McDonald distinguished Herst on the facts, and noted that the McDonald decision was dependent upon the evidence before the jury in that case. (Id. at pp. 883-884.)

Citing McDonald , defendant contends “the trial court erred in instructing the jury that the defendant had the burden of raising a reasonable doubt regarding his partition ratio.”

*Supp. 17 Respondent asks that we distinguish McDonald and follow Herst .

McDonald distinguished Herst, supra, 197 Cal.App.3d Supp. 1 , because the Herst court was able to presume that a person’s partition ratio is constant and can be measured by the defendant. ( McDonald, supra, 206 Cal.App.3d at pp. 881-883.) However, in McDonald, defendant’s criminalist, Henry Greenberg, testified that a person’s partition ratio varies over time, and therefore while a person’s ratio could be tested, it would be speculative to say the ratio was the same at the time of the test after arrest as it would be at the time of a test of his ratio some weeks later. He also opined it would be costly to conduct a series of tests to determine the variability of a person’s partition ratio. (McDonald, supra, at p. 880 .) The evidence in McDonald therefore precluded application of the rule of convenience relied on in Herst. (Id. at pp. 883-884.)

In this case, the defendant said he stopped drinking four hours before the test. This put him in the postabsorptive phase at the time of the test. Though there was testimony about the absorptive phase (when the partition ratio can vary widely), the testimony was more focused on the postabsorptive phase than is reflected in McDonald . Mr. Greenberg, the criminalist quoted in McDonald , testified on behalf of defendant in this case. Here, however, he did not testify that a person’s partition ratio during the postabsorptive phase varies over time, and he did not testify that it would be costly to determine the defendant’s partition ratio. Mr. Greenberg did testify that he has tested individuals to determine their partition ratios, and he could have tested defendant to ascertain his partition ratio but he did not. (See appen. 1.)

The McDonald case is therefore factually distinguishable by its own analysis. (P eople v. McDonald, supra, 206 Cal.App.3d at pp. 881, 883.) McDonald distinguished Herst factually rather than disapproving it. As respondent contends, the evidence in the present case is more similar to Herst than to McDonald . However, we conclude that the propriety of the challenged instruction does not depend upon evidentiary differences between cases.

We cannot take judicial notice that a person’s postabsorptive partition ratio does, or does not, vary significantly within a few days or weeks. (Compare McDonald, supra, 206 Cal.App.3d at p. 884 .) That factual issue need not be resolved to determine the propriety of the instruction.

The Legislature could have prohibited driving a vehicle when a subsequent breath test shows a blood-alcohol level of 0.10 percent or more. It did not. (Burg v. Municipal Court (1983) 35 Cal.3d 257, 266, fn. 10 [ 198 Cal.Rptr. 145 , 673 P.2d 732 ].) Other evidence is admissible to *Supp. 18 challenge or support the test’s results, and to establish defendant’s blood-alcohol level and conditional the time of driving. (Ibid.; People v. Randolph (1989) 213 Cal.App.3d Supp. 1 [ 262 Cal.Rptr. 378 ].) Evidence regarding test accuracy is admissible regardless of whether in fact a defendant can determine his ratio on the date of the arrest by means of tests conducted hours, days, or weeks after his arrest. Evidence of the range of individual blood-breath ratios cannot be excluded, even in the absence of evidence of a defendant’s personal ratio. Consonant with prior unpublished opinions of this court, evidence of the range of partition ratios was received in this case.

Vehicle Code section 23155 does provide a presumption (permissible inference) if the trier of fact finds beyond, a reasonable doubt that a defendant’s actual blood-alcohol level at the time of the test was 0.10 percent or more. A finder of fact may infer the actual blood-alcohol level at the time of the test from the breath test results, but the statute does not provide a presumption that the test results are accurate reflections of the actual blood-alcohol level.

Regardless of any theoretical foundation for a rule of convenience, we do not believe it is appropriate to judicially create a presumption of accuracy of the breath test results. Similarly, we do not believe it is appropriate to judicially create a presumption regarding facts that might theoretically affect the accuracy of the test’s results, such as an individual’s blood-breath ratio.

The California Code of Regulations requires breath test instruments to use a partition ratio of 2,100 to 1. The jury may, but is not required to, infer that a defendant’s ratio is 2,100 to 1. The prosecution is required to prove each element of the crime charged beyond a reasonable doubt. The defendant’s partition ratio is not an element of the crime, and the prosecution is not required to prove the defendant’s partition ratio as such. In determining defendant’s blood-alcohol level at the time of a test, and in determining whether defendant is guilty of the crimes charged, the jury may consider all the evidence received during the trial. (Burg, supra, 35 Cal.3d at 266, fn. 10 ; People v. Randolph, supra, 213 Cal.App.3d Supp. 1 .)

Citing Herst, supra, 197 Cal.App.3d Supp. 1 , respondent contends the instruction did not shift the burden of proof for an element of the offense, but rather merely placed the burden on the defendant to prove an affirmative defense—that his partition ratio deviated from the norm. (6) The burden of proof on an affirmative defense may be allocated differently from the burden for elements of the offense. (See 1 Witkin, Cal. Evidence (3d ed. 1986) §§ 166-169, pp. 141-143.) It is true, as stated in Herst , that a defendant’s personal partition ratio is not an element of the *Supp. 19 crime. However, that does not automatically convert it to an affirmative defense on which defendant has the burden of proof. (See Witkin, supra-, People v. Tewksbury (1976) 15 Cal.3d 953 [ 127 Cal.Rptr. 135 , 544 P.2d 1335 ]; and Wilson, Shifting Burdens in Criminal Law: A Burden on Due Process (1981) 8 Hastings Const. L.Q. 731 , 751.) Disproving or casting doubt on an element of a charged offense is not an affirmative defense like the defense of necessity or the defense of self-defense.

Herst appears to assume that only evidence of defendant’s ratio as determined by a test can rebut an inference that his ratio is 2,100 to 1. There may be evidence that the defendant consumed less alcohol than implied by the result of a breath test. There may be evidence that his driving, behavior, and performance on field sobriety tests were consistent with sobriety rather than the degree of influence implied by the result of a breath test. We see no reason why he cannot seek to persuade the trier of fact that, rather than being incredible because of conflict with a breath test, the exculpatory evidence is believable and he may have a blood-breath ratio sufficiently below 2,100 to 1 to be reconcilable with the breath test results.

We find no basis for a presumption such as implied in Herst . We similarly find no basis for placing on defendant the burden of raising a reasonable doubt regarding his partition ratio. We therefore conclude that the instruction the trial court gave in this case was erroneous.

Prejudice

The analysis for prejudice from an erroneous instruction regarding the burden of proof is particularly delicate because such an instruction may “conflict with the overriding presumption of innocence with which the law endows the accused and which extends to every element of the crime.” (Morissette v. United States (1952) 342 U.S. 246, 275 [ 96 L.Ed. 288, 307 , 72 S.Ct. 240 ].) It may also invade the fact-finding function of the jury. (United States v. United States Gypsum Co. (1978) 438 U.S. 422, 446 [ 57 L.Ed.2d 854, 875 , 98 S.Ct. 2864 ].) However, even an instruction which conclusively presumes a disputed element of the crime charged is subject to harmless error analysis. (Carella v. California (1989) 491 U.S. 263 [ 105 L.Ed.2d 218 , 109 S.Ct. 2419 ],

The instruction in the present case did not state a conclusive presumption, and did not apply to an element of the offense.

The trial court gave standard instructions regarding the burden of proof, circumstantial evidence, and permissible inferences. Under some circumstances, the error might be harmless. However, in this case, the erroneous *Supp. 20 instruction was given to the jury separately in response to a specific inquiry from the jury asking which side had the burden of proof regarding the partition ratio. The inquiry apparently was stimulated by the way the case was presented and argued.

Since the defendant’s personal partition ratio is not an element of either crime charged, neither party had any burden of proof of that fact per se. If conviction of a charge depends on the accuracy of the breath test results, and if the finding on that issue depends on the defendant’s personal ratio (or at least that the defendant’s ratio is no lower than a certain ratio), then the burden of proof of that fact is on the prosecution. The burden is not on the defendant to raise a reasonable doubt as to his personal partition ratio. Under the circumstances in which the instruction was given, it was susceptible to the interpretation that if the defendant failed to raise a reasonable doubt as to whether his personal blood-breath ratio was average, then he should be convicted.

Thus, even though the case presented few evidentiary conflicts, under the circumstances we cannot conclude that the error was harmless. (People v. Watson (1956) 46 Cal.2d 818, 836 [ 299 P.2d 243 ].)

The judgment is reversed and the case is remanded for retrial.

Jones, P. J., and Lane, J., concurred.

*Supp. 21 Appendix 1

We requested and obtained from the municipal court its entire file and tape recordings of the trial, to guard against any possible inadvertent but significant omission from the statement on appeal. During his direct examination, Mr. Greenberg testified: “Q. We talked just a little bit about the partition ratio. Have you ever conducted any studies yourself to determine this partition ratio? “A. Yes. “Q. And would you explain to the jury what sort of a test that you yourself have conducted. “A. Basically what I’ve done is I’ve interviewed individuals, found out what their drinking habits or drinking patterns are and given them quantities of alcohol, brought their alcohol levels up to anywhere between a .08 and .15, .16 on some occasions, measured blood-alcohol levels by blood and by breath using the Intoxilyzer 4011A at various stages while the alcohol level was rising, at the maximum or peak level, and in the post-absorptive phase. I calculated ratios, blood-to-breath ratios, and if the set of circumstances for that particular individual warranted, the information was then used in court.” During cross-examination, Mr. Greenberg testified: “Q. All right, did you say that you yourself have measured the partition ratio of individuals? “A. Yes. “Q. Have you measured the partition ratio of Anthony Cortes? “A. No, I have not. “Q. Do you have any reason to believe that Anthony Cortes’ partition ratio is lower than the 2100 to 1 partition ratio which is mandated for the Intoxilyzer machine to be set at? “A. I would expect that if he were tested he would fall within the normal—what I would consider the normal range in the post-absorptive phase, so I would expect him to fall between approximately 1600 and 3050, 3100. “Q. I don’t think that answered my question. Do you have any reason to believe that his partition ratio is lower than 2100 to 1. “A. I have no way of knowing if it would be lower or higher. “Q. Well, you do have a way of knowing, don’t you? Couldn’t you test Mr. Cortes and find out what his partition ratio is? “A. Yes. “Q. But you didn’t in this case? “A. No, I did not. “Q. Wouldn’t that have given you a more concrete basis for challenging the Intoxilyzer result in this case if you had taken Mr. Cortes into your laboratory and found out exactly what his partition ratio is? “A. The more information available to an individual to a scientist, the better—the more conclusive—opinion or estimation that can be drawn, so I would say yes.”