Opinion

In this prosecution for driving while having a proscribed alcohol concentration in blood or breath, the trial court excluded expert witness testimony challenging the reliability of breath-alcohol testing machines. Reviewing the ensuing conviction, the Court of Appeal determined that the trial court erred, and reversed for a new trial. We conclude that the trial court properly excluded the challenged expert testimony with respect to the charge of driving while having a proscribed alcohol concentration. Accordingly, we reverse the judgment of the appellate court and affirm the conviction.

I. Facts and procedure

On December 22, 2007, at approximately 2:30 a.m., Sergeant Richard W. Berg of the California Highway Patrol (CHP) observed defendant Terry Vangelder, who was then 50 years of age, driving a high-performance pickup truck more than 125 miles per hour on Highway 163 in San Diego County. Sergeant Berg followed defendant for more than five miles and eventually caught up with him, noticing that he slowed to 100 miles per hour when he approached other traffic and was not weaving outside his lane. After traveling an additional mile and one-half, Berg activated his patrol vehicle’s red lights. Defendant rapidly decelerated, pulled over, and, as directed, proceeded to a wider spot on the shoulder.

Defendant provided his license and registration and said, “I was just screwing around.” Berg detected an odor of alcohol and noted that defendant’s eyes were red and watery. Defendant admitted he had consumed two glasses of wine. Berg called for backup and transferred the matter to two other CHP officers, including Gerald Guzman, who arrived at 2:58 a.m.

Officer Guzman gave defendant field sobriety tests, in which defendant exhibited few signs of impairment. Defendant told Guzman he had three glasses of wine at dinner. Guzman thought defendant smelled of alcohol and noticed that his eyes were red and glassy. Soon thereafter, at approximately 3:10 a.m., defendant consented to two in-field preliminary alcohol screening breath tests using an Intoximeter Aleo-Sensor IV. Such a test is “preliminary” in the sense that it is employed—only with the driver’s actual consent—prior to any arrest, in order to assist an investigating officer in determining whether to arrest the driver. Standard testing protocol required that a subject be observed continuously for 15 minutes before the test was administered, in order to make sure that the subject had not during that time ingested alcohol or done anything else that might compromise the test (see Cal. Code Regs., tit. 17, § 1219.3, 3d sentence), but Officer Guzman waited and observed for only nine or 10 minutes, reasoning that defendant had generally been under earlier surveillance for many minutes by Sergeant Berg. Guzman testified that defendant displayed an alcohol concentration of 0.095 percent on the first test, and 0.086 percent on the second test, which was administered two minutes later.

Based on his observations and the preliminary breath tests, Guzman believed that defendant was under the influence of alcohol, arrested him, and transported him to the county jail. There, defendant was subject to additional chemical testing under the implied consent law, Vehicle Code section 23612 (subsequent statutory citations are to this code unless otherwise indicated), which provides for testing of blood, breath, or urine. He elected breath testing, which was conducted using an Intoximeter EC/IR. The first breath test from that device, taken at 3:37 a.m., registered a level that revealed an 0.08 percent alcohol level, and the second test, taken two minutes later, produced the same result.

Finally, defendant additionally consented to a blood test, taken at 3:52 a.m. These samples, tested at a later date, showed alcohol percentage levels of 0.088 and 0.087, respectively.

The San Diego City Attorney’s Office filed a misdemeanor complaint charging defendant with (1) driving under the influence of alcohol in violation of section 23152, subdivision (a) (hereafter section 23152(a))— sometimes referred to as the “generic offense”; and (2) driving with a blood-alcohol concentration of 0.08 percent or more, in violation of section 23152, subdivision (b) (hereafter section 23152(b))—sometimes referred to as the “per se offense.”

Prior to trial, defendant filed motions to (1) exclude evidence of the PAS tests to establish blood-alcohol content, asserting they were not conducted in accordance with applicable regulations and were unreliable and (2) allow evidence concerning “partition ratio variability” (described post, pt. II.A.3.) with regard to the generic offense. The People filed opposing motions. The trial court deferred ruling on the PAS and partition ratio evidence issues, and the matter proceeded to trial.

A. The People’s evidence

CHP Officer Brandon Garland, who was responsible for maintaining and calibrating the agency’s PAS breath-testing devices, testified regarding the requirements for successfully conducting a PAS test. Garland’s records showed that the particular Intoximeter Aleo-Sensor IV PAS machine used to analyze defendant’s breath at the scene of the vehicle stop had been tested earlier that week and had been found to be operating within acceptable limits of error.

Marissa Ochoa, a criminologist at the San Diego County Sheriff’s Regional Crime Laboratory, testified that both the particular Intoximeter EC/IR breath-testing machine and the blood sampling device used to test defendant at the county jail were in proper working order when used to take and measure samples from defendant, and that the Intoximeter EC/IR breath-testing machine in question has an operational range within the accepted limits of error (see ante, fn. 4). Ochoa further testified regarding the effect of alcohol on the human body. She was asked to assume that a healthy male weighing 200 pounds had consumed three glasses of wine with a hearty meal between 8:00 and 9:00 p.m., was pulled over while driving at 2:45 a.m., and thereafter at the jailhouse produced a breath test result of 0.08 percent at 3:37 a.m. and a similar blood test result 15 minutes later. She concluded that under normal bodily processing of alcohol, such a “person[’]s alcohol level [would] be 0.09, approximately, at the time of driving.” On cross-examination, Ochoa explained that if an average healthy male weighing 200 pounds had ingested alcohol only during a dinner between 8:00 and 9:00 p.m., he would have needed to consume approximately 11 drinks (each drink defined as a 12-ounce beer, a 4- to 6-ounce glass of wine, or a 1- to 1.25-ounce serving of hard liquor) in order to have a 0.08 percent blood-alcohol level seven and one-half hours later.

B. Defendant’s evidence

1. Initial testimony by Dr. Hlastala concerning asserted unreliability of breath-testing machines in light of alcohol that is “picked up” during inhalation before reaching the deep lungs

Defendant called Dr. Michael P. Hlastala, professor of medicine, physiology, biophysics, and bioengineering at the University of Washington, to testify concerning the effects of alcohol on the respiratory system and the reliability of breath-testing machines. Hlastala explained that his field of study focuses on “the physiology of the human body, which means I deal with the lungs, blood and the way that substances move around the body . . . and one of those is alcohol, [which] I’ve been studying ... for about 25 years, in terms of the physiology of alcohol and also the way alcohol is measured in testing procedures.” He related that he had written more than 400 articles, including approximately 170 peer-reviewed professional articles, and a textbook. He explained that he had testified in approximately 30 states, and served on committees of the National Institutes of Health. The People did not dispute Hlastala’s expertise.

Defense counsel asked Dr. Hlastala to assume that both a preliminary “roadside breath test” (i.e., PAS) machine and an “ECZRI” breath-testing machine are “working perfectly, and they’ve been properly done. Do they provide a scientifically accurate test?” Hlastala replied, “Well, they don’t.” (Italics added.)

After a sidebar conference, Dr. Hlastala continued by explaining “the way we take oxygen and put it into the body” and “the way we eliminate carbon dioxide” from the body. He used a pen to create a diagram and explained: “Starting from the throat, the windpipe is called the ‘trachea.’ The trachea comes down ... to just about the heart, . . . and that splits off into a left and right side, and that splitting goes on . . . about 20 times before it gets into the air [sacs]” deep in the lungs, known as the alveolar sacs. “There are about 300 million of these [alveolar sacs] in every valve. They are . . . very tiny. . . . There’s blood vessels around that and this is where a lot of the action takes place. If we bring outside air with oxygen in it, the oxygen goes into the blood, . . . and then it gets . . . metabolized and that provides energy for us.”

Thereafter, Dr. Hlastala began to testify that even if breath-testing machines operate as designed, they do not perform a scientifically reliable test, because some alcohol present in mucus membranes and bronchial vessels in the upper airways absorbs into the breath before it finally reaches the alveolar sacs in the deep areas of the lungs. Specifically, he stated: “Now, when the breath test was developed in the 1950s, it was understood that if there was alcohol in the blood, that some of it would get out into ... the [alveolar] air sac[]s and ... the idea ... of a breath test” is to take this air originating in the alveolar sacs “and breathe it out, and then measure it. . . with the breath test instrument. . . . [¶] The concept is that this alcohol in this [sampled] air, [is] equal to what’s down in here [in the alveolar sacs], [and] hence [it’s] related to whatever’s in the blood. We know, now, that it’s not quite that simple because alcohol is quite soluble, it goes into water quite easily. And we have, in the airway, a lot of mucus and water and that mucus lining in the airway plays an important role in protecting us from particles and things we inhale[,] goes on to this mucus, then comes out to the mouth.”

Dr. Hlastala continued: “But if we have alcohol, there are little blood vessels that come along here, and these blood vessels, those are called ‘bronchial vessels.’ And so they bring alcohol so there’s a lot of alcohol if you have alcohol in your bloodstream. Now, what happens is if we inhale and we pick up alcohol from this mucus ... by the time we get down to this [alveolar] air sac[f], it’s already filled up and saturated.” (Italics added.)

In other words, Dr. Hlastala informed the jury that although breath-testing machines are designed to sample and analyze the concentration of alcohol contained in alveolar, deep lung air, they fail to do so. At this point the prosecutor objected to Dr. Hlastala’s testimony. The trial court excused the jury until later in the afternoon and conducted a hearing regarding the admissibility of the initial and further proposed testimony.

2. Hearing concerning additional proposed testimony by Dr. Hlastala that other physiological factors make breath-testing machines unreliable

Outside the jury’s presence, the trial court questioned whether testimony by Dr. Hlastala would constitute evidence concerning “partition ratio variability,” which we had earlier held in Bransford, supra, 8 CalAth 885, to be inadmissible in prosecutions under the per se statute, section '23152(b). Thereafter, under questioning by counsel and the court, Dr. Hlastala testified that various other factors can cause a breath-alcohol test to be scientifically inaccurate. He identified those other factors as the pattern of breathing (speed and depth of inhalation and exhalation); body and breath temperature; and hematocrit level (ratio of red blood cells to total blood volume).

On cross-examination the prosecutor asked: “Doctor, in this case, we have a result from an EC/IR [breath-testing machine] of 0.08. In your opinion, what does that 0.08 measure?” Dr. Hlastala responded: “It’s measuring the alcohol concentration of the breath that is delivered to the machine.” The prosecutor asked, “Is there any reason to believe, in this case, that [reading] was inaccurate?” Dr. Hlastala responded, “It’s an inaccuracy in how ... the alcohol comes out of the mouth” because “from time to time, a person . . . can deliver different amounts of alcohol through the mouth . . . through the breath.”

The prosecutor continued: “So, in your opinion, breath tests are inherently inaccurate as a measure of how much alcohol a person has in them?” (Italics added.) Dr. Hlastala answered: “They are. And primarily because the basic assumption that all of the manufacturers have used is that the breath that [is] measured is directly related to water in the lungs, which is directly related to what’s in the blood. And in recent years, we’ve learned that, in fact, that’s not the case.” (Italics added.)

The prosecutor followed up: “So that . . . wouldn’t mean that the machine has, in a given sample, not measured it accurately?” Dr. Hlastala responded, “No. I’m assuming that the machine is working accurately.” He elaborated by stating that breath-testing machines cannot reliably reflect the concentration of alcohol in blood because they constitute “an indirect test” of blood alcohol. He explained that the best sample of blood would be that found in the brain, and that a sample from venous blood “is the next best thing.” As an alternative to such a blood sample, “the breath is the next best thing. [And then] [t]he odor of alcohol is the next best thing. They’re all remote, the more remote the more variable.”

At this point the court reviewed with Dr. Hlastala the other factors that, according to his proposed testimony, render breath testing scientifically unreliable. In addition to the factors that Hlastala had focused on earlier in the hearing outside the jury’s presence—pattern of breathing (speed and depth of inhalation), body and breath temperature, and hematocrit level (ratio of red blood cells to total blood volume)—the witness stated that additional relevant factors include sex (explaining that “[w]omen have a smaller lung, therefore a higher breath test relative to that blood”), and “medical condition,” such as “lung diseases.”

The court observed that California law, in correlating breath alcohol with blood alcohol, applies “what’s known as the ‘standard partition ratio,’ which I understand is grams of alcohol in 210 liters [of breath] equals grams of alcohol per 100 milliliters of blood”—and the court asked the witness if he believed that “the [true] ratio of one person might be different [from the ratio in] someone else?” Dr. Hlastala responded: “That’s correct. . . . [S]ome people would be different.” The court pressed, asking: “Now, . . . basically what you’re saying here as to why this [breath-testing machine] test isn’t scientifically reliable, [is] because it automatically applies this standard partition ratio?” Dr. Hlastala replied: “No, not really. I’m not talking about the partition ratio. I’m talking about factors that influence the breath [sample], breath-alcohol [sample], and I’m not talking about comparing it to blood.”

The court continued, “[I]n this particular case, are you prepared to state that the breath sample of defendant as being 0.08 would be . . . overstating or understating?” Dr. Hlastala responded that it could be either, and that he had no opinion whether defendant’s breath test result of 0.08 percent was actually (or even probably) overstated or understated.

The court asked Dr. Hlastala to step outside and then entertained further argument from counsel. The prosecutor observed that section 23152(b) criminalizes driving with a proscribed blood-alcohol level or a certain proscribed breath-alcohol level as measured by a properly calibrated and administered breath-testing machine—and he asserted that “the question of whether the breath [sample and result] accurately reflects anything else is irrelevant.” The prosecutor argued that in this context a defendant should not be allowed to present expert testimony that “there was something wrong with the way that the State of California has drafted its laws.” The prosecutor moved to “strike any testimony that the doctor gave with regard to the diagram” that he had drawn during his testimony before the jury, and asked “that the diagram be removed from the view of the jury. . . . Because that’s all about how . . . alcohol came from the blood and into the trachea, and how . . . alcohol [is brought in breath] from the trachea and other sources.”

Although the court appeared to agree with the prosecutor that “at some point” it would instruct the jury not to consider Dr. Hlastala’s testimony that even if breath-testing machines are working properly, their samples and results are not scientifically reliable due to physiological factors, the court was unwilling to so instruct the jury immediately. Instead, the court said it would “think about it”—but that “[a]t a minimum, I’m going to instruct counsel not to argue that.”

Eventually the trial court ruled—over defendant’s objections—that Dr. Hlastala’s proposed testimony was speculative and did not materially differ from evidence concerning partition ratio variability, which we had held in Bransford, supra, 8 Cal.4th 885 (discussed post, pt. II.B.2.), to be inadmissible in prosecutions under the per se statute, section 23152(b). Accordingly, the trial court ruled that although it would allow Dr. Hlastala to resume testifying before the jury, there were to be “[n]o questions to this expert, which will solicit any testimony by him” that, as a general matter, properly working breath-testing machines do not sample air as they are designed to and do not produce reliable results because of the factors mentioned in his prior (prehearing) testimony before the jury or in his subsequent testimony outside the jury’s presence. The court made clear, however, that defendant remained free to argue, and present evidence, that the particular machines used in this case malfunctioned, were not properly calibrated, or that they were not properly employed. Moreover, the court reiterated, Dr. Hlastala could be questioned about the contaminating effect of “mouth alcohol” (alcohol that stays in the mouth after ingestion, and that dissolves into mouth tissues) on the reliability of a breath test sample and result.

3. Resumed limited testimony by Dr. Hlastala

Thereafter, Dr. Hlastala resumed the stand and testified that the presence of mouth alcohol can cause a “false positive test,” meaning a higher result than would be obtained otherwise. Dr. Hlastala also testified that “retrograde extrapolation”—estimating a blood-alcohol level that existed at an earlier time based on a reading at a known later time, as the People’s witness, Ochoa, had undertaken—was “difficult” to do “over long periods of time.” He did not testify that the particular breath-testing machines used in this case malfunctioned, or that they were improperly calibrated.

4. Testimony by defendant and his son

Defendant’s 16-year-old son Wesley testified that defendant drank two or three glasses of wine between 8:00 and 9:00 p.m. during dinner at a restaurant. Thereafter, Wesley related, they went for a night hike and returned home about 2:00 a.m., when defendant drank a beer and Wesley went to bed.

Defendant testified that at approximately 2:00 a.m. he drank one beer, and then went out again, driving his vehicle on Interstate 15 and Highway 163. He explained that the night was moonlit, traffic was light, and “I did a burst of speed. . . . And there was no question, I exceeded 100 miles per hour.” Defendant testified that when he was informed by the officer of the results of his breath tests, “I, quite frankly, did not believe,. . . that that’s what I was at. So I said I wanted to submit to a blood test. I didn’t think it was accurate. And the reason I say that is because I knew how much I had to drink. I knew what the effects were. I mean, obviously you have a beer you might feel the tingle, but it did not affect my motor skills. It didn’t affect my burst of speed. That’s just who I am. I would have done that whether or not I had that beer or not. That’s just something that would have happened.”

C. Jury deliberations and verdict

During deliberations, the jury inquired whether it was allowed to find a defendant guilty of driving with a blood-alcohol content above 0.08 percent (the per se charge, § 23152(b)), and yet not guilty of driving while under the influence (the generic charge, § 23152(a)). The court answered, “Yes.” The jury ultimately was unable to reach a verdict on the generic charge, but found defendant guilty of driving with 0.08 percent or more of alcohol in his blood. The jury also found defendant guilty of driving in excess of 100 miles per hour, in violation of section 22348, subdivision (b). The trial court suspended imposition of sentence on the per se charge for five years; imposed but then stayed a fine of nearly $2,000 and standard educational conditions relating to that charge; and imposed a fine of nearly $2,000 on the speeding count. The generic count was set for retrial, but the trial date was continued pending appeal.

Defendant filed a notice of appeal to the appellate division of the superior court. The appellate division denied the appeal, but the Court of Appeal granted defendant’s application for certification to transfer the case. (Cal. Rules of Court, rule 8.1005.) Ultimately, the appellate court rejected the trial court’s rulings on Dr. Hlastala’s testimony and proposed testimony, specifically disagreeing with that court’s conclusion that the testimony, insofar as it related to the per se charge, was barred as inadmissible evidence concerning partition ratio variability under Bransford, supra, 8 Cal.4th 885 . The Court of Appeal found prejudicial error, requiring reversal of the section 23152(b) per se conviction. We granted the People’s petition for review.

II. Contentions and relevant background principles

The People assert that the appellate court erred both in overturning the trial court’s exclusion of the prehearing jury testimony and subsequently proffered testimony of Dr. Hlastala, and also in finding the alleged errors to be prejudicial. Defendant, for his part, asserts the trial court erred in excluding the proffered testimony (and argument concerning it) that breath-testing machines are unreliable because even when they operate and are employed as designed, they do not perform a scientifically reliable test. Specifically, as further described post, in part III., defendant presents two general contentions: First, he asserts the trial court erroneously precluded him from presenting as a defense the argument that breath-testing machines are unreliable due to alcohol that is “picked up” during inhalation while traveling past mucus membranes lining the airway before breath reaches the alveolar, deep lung regions—and that, accordingly, the machines fail to sample alcohol contained within alveolar or deep lung air, as required by California Code of Regulations, title 17, section 1219.3. Second, defendant asserts the trial court erred in barring Dr. Hlastala’s additional proposed testimony, adduced at the hearing outside the presence of the jury, that other physiological factors, such as pattern of breathing (speed and depth of exhalation), body and breath temperature, and hematocrit level (ratio of red blood cells to total blood volume), also make breath testing unreliable. In both respects, defendant asserts, the trial court erred in equating Dr. Hlastala’s testimony with evidence concerning partition ratio variability, which defendant concedes would be inadmissible in a section 23152(b) per se prosecution under Bransford, supra, 8 Cal.4th 885 .

Before addressing these contentions in part III., it is useful to review (A) relevant scientific principles, assumptions, and corresponding regulations and (B) relevant legislation and related case law.

A. Scientific principles, assumptions, and corresponding regulations

After ingestion and absorption through the stomach walls and the intestines, ethyl alcohol enters the blood and eventually travels via the carotid arteries to the brain, where it causes intoxication and resulting mental and physical impairment. (McNeal, supra, 46 Cal.4th at pp. 1190-1191; State v. Chun (2008) 194 N.J. 54 [ 943 A.2d 114, 126 ] (Chun); see generally Mason & Dubowski, Breath-Alcohol Analysis: Uses, Methods, and Some Forensic Problems—Review and Opinion (1976) 21 J. Forensic Sciences 9 (hereafter Mason and Dubowski, Breath-Alcohol Analysis).) At the same time that absorption of alcohol occurs, elimination also commences through excretion and metabolization. “When a person’s body is absorbing alcohol faster than he or she is eliminating it, the concentration of alcohol in the blood will continue to rise. . . . The concentration will reach its peak, and it will achieve a plateau, at the time when elimination and absorption are occurring at about the same rate. [][] [Thereafter,] [w]hen the person . . . slows down ingestion to the point where the body is eliminating alcohol more quickly than absorbing it, the body enters what has generally been referred to as the post-absorptive phase. During this period of time, the concentration of alcohol in the blood decreases.” (Chun, supra, 943 A.2d at p. 127 .)

1. Sources for measuring alcohol concentration

Although blood in the brain itself—or in the carotid arteries leading to the brain—would be the best source from which to test for the presence of impairing alcohol, as a practical matter it is impossible to acquire such samples. (McNeal, supra, 46 Cal.4th at p. 1191; State v. Downie (1990) 117 N.J. 450 [ 569 A.2d 242, 246 ] (Downie).) Instead, tests focus on the next best sources: samples from venous (or capillary) blood, from breath, or from urine. (See, e.g., § 23612 [implied consent for blood, breath, or urine testing]; Cal. Code Regs., tit. 17, §§ 1219.1 [blood collection], 1219.2 [urine collection], 1219.3 [breath collection].)

Whereas testing of venous blood directly measures the concentration of alcohol in a person’s blood, breath and urine testing provide indirect measures for determining the concentration of alcohol in blood. The New Jersey Supreme Court has described the generally accepted scientific understanding concerning how alcohol becomes subsumed into breath: “Alcohol passes” from the pulmonary blood vessels “into the lungs, through the walls of the air sacs, called alveoli. As it does so, it mixes with the air that the person has inhaled. When the person exhales, alcohol passes out of the body as part of the breath, [f] An individual’s breathing pattern can influence the amount of alcohol that appears in any particular breath. In addition, the amount of alcohol in the breath sample represented by a single act of exhalation will vary from the beginning to the end. This is because the breath actually comes from different parts of the body, from the mouth to the deepest part of the lungs. Except for the possible interference that would occur if the test subject had ingested alcohol so recently that residual mouth alcohol were captured, the first part of the breath comes from the mouth and throat where there is little contact with the alcohol passing through the alveoli. However, as the person continues to exhale, the expelled air comes from deeper in the respiratory system, where it contains alcohol that more closely represents the amount passing through the lungs from the circulating blood.” (Chun, supra, 943 A.2d at p. 127 , italics added; accord, McNeal, supra, 46 Cal.4th at p. 1191 [“When a subject blows into a breath-testing machine, the device measures the amount of alcohol vapor expelled into alveolar spaces deep in the lungs.”].)

2. “Henry’s Law” and the target breath sample

Exhaled breath emanating from the region of tiny alveolar sacs deep in the lungs is deemed to reflect “that portion of the expired breath which is in equihbrium with respect to alcohol with the immediately adjacent pulmonary blood.” (Cal. Code Regs., tit. 17, § 1215.1, subd. (m) [defining alveolar].) A sample of this type of breath is assumed to satisfy the conditions of Henry’s Law, a scientific principle holding that the concentration of a volatile substance (in this instance ethyl alcohol) dissolved in liquid (in this instance blood) in a closed environment (in which factors such as pressure or temperature are fixed) is directly proportional to the concentration of that substance in the air next to that liquid. (Downie, supra, 569 A.2d at p. 246 ; see generally Annot. (1991) 90 A.L.R.4th 155 , 159-160; 2 Erwin, Defense of Drunk Driving Cases (3d ed. 2013) § 18.01[2][a], pp. 18-5 to 18-7 (hereafter Defense of Drunk Driving Cases) [discussing Henry’s Law].)

Breath-testing machines are designed to measure breath-alcohol levels in light of the conditions under which Henry’s Law operates in the alveolar regions deep in the lungs where the gaseous exchange between pulmonary blood and inhaled air occurs. (See generally 2 Defense of Drunk Driving Cases, supra, § 18.01[2][a], p. 18-7; see also Flores et al., Breath Alcohol Sampling Simulator (BASS) for Qualification Testing of Breath Alcohol Measurement Devices (1981) U.S. Dept. of Commerce, National Bureau of Standards Special Publication 480-41, p. 3 [“the alveoli-blood interface is the primary locus for active gas exchange between blood and breath . . .”] (hereafter BASS for Qualification Testing of Breath Alcohol Measurement Devices).) Thus, a target breath sample often is described, as in the relevant federal regulation and in the statutes or regulations of many states, as “deep lung” or “end-expiratory” air or breath—in other words, the last portion of the exhalation of a deep breath. Some state provisions describe the same target sample as “alveolar in composition,” “substantially . . . expired alveolar air,” or employ a variation on that terminology. Other state provisions combine this terminology, calling for samples of “alveolar or deep lung air,” “deep lung (alveolar) air,” “deep lung breath (substantially alveolar in composition),” or a sample that is “essentially alveolar or end expiratory in composition.” The variation employed by California and a few other states calls for a sample of “expired breath which is essentially alveolar in composition.” (Cal. Code Regs., tit. 17, § 1219.3.)

However phrased, the relevant federal and state regulations all require a specimen originating from deep within the lungs—a sample that is considered to most closely reflect the concentration of alcohol passing from the circulating blood into the alveolar sacs. Correspondingly, breath-testing machines have long been designed to capture and measure such a sample, obtained from the last portion of the expired breath. (Dubowski, The Technology of Breath-Alcohol Analysis (1991) U.S. Dept, of Health & Human Services, pub. No. (ADM)92-1728, pp. 5-6.) Subjects are instructed to take and expel a deep breath. (See, e.g., Chun, supra, 943 A.2d at p. 129 .) Models like the Intoximeter EC/IR used in this case are designed to indicate whether an adequate sample of end-expiratory breath has been provided—and to prompt the operator to secure a subsequent sample as necessary. (See, e.g., 2 Defense of Drunk Driving Cases, supra, § 18.05[l][b], p. 18-141; see also id., § 19.05[3], pp. 19-10 to 19-11 [describing the “sampling requirements” of the Intoximeter EC/IR used in the present case]; cf. Chun, supra, 943 A.2d at pp. 129-131 [describing procedures used for a related model].)

3. Scientific analysis of breath samples—(a) the blood-alcohol: breath-alcohol partition ratio, and (b) “partition ratio variability”

Scientific analysis of breath samples to determine the concentration of alcohol in blood is premised on a conversion methodology utilizing what is known as the “blood-alcohol:breath-alcohol partition ratio.” The standard partition ratio long used in California (and indeed, throughout the U.S. and most other countries) is legislatively set at 2,100 to one—meaning that the amount of alcohol in 2,100 milliliters of a breath sample is deemed to correspond to the amount of alcohol in one milliliter of blood. (McNeal, supra, 46 Cal.4th at pp. 1188, 1191, citing Cal. Code Regs., tit. 17, § 1220.4, subd. (f); see 2 Defense of Drunk Driving Cases, supra, § 18.01[2][b], p. 18-7 [“[I]t is assumed that the concentration of alcohol in blood is 2100 times that in the breath or that it takes 2100 parts of breath by volume to have the same amount of alcohol as is in one part of blood by volume.”].)

This 2,100-to-one partition ratio (sometimes stated in terms of the amount of alcohol in 210 liters of breath and 100 milliliters of blood) is a scientifically agreed-upon construct arrived at by use of what the parties in their briefs refer to as “black box methodology”—in other words, performance studies that measure and correlate both an individual’s blood sample and that same individual’s simultaneously collected breath sample—as analyzed by a properly functioning breath-testing machine. (See Downie, supra, 569 A.2d at p. 248 [describing studies of “paired venous blood/breath samples”]; see also State v. Hanks (2001) 172 Vt. 93 [ 772 A.2d 1087, 1089 ] [“a conversion rate of 2100:1 [is] an assumed blood-breath ratio, which represents the relationship between the number of alcohol molecules in the bloodstream to the number present in the breath when both substances are tested simultaneously”]; see generally Mason & Dubowski, Breath-Alcohol Analysis, supra, 21 J. Forensic Sciences No. 1, at p. 16 et seq.)

It is well documented by scientists and accepted by courts that true partition ratios vary both between individuals and within individuals over time. This “partition ratio variability” is known to result from various physiological factors including breathing patterns (speed and depth of exhalation), body temperature, hematocrit level (ratio of red blood cells to total blood volume), sex, and each person’s medical health. (See McNeal, supra, 46 Cal.4th at p. 1191 [listing, as variables affecting the actual ratio of an individual’s breath-alcohol concentration to blood-alcohol concentration," ‘body temperature, atmospheric pressure, medical conditions, sex, . . . the precision of the measuring device,’ ” and “hematocrit level and elapsed time between drinking and breath-alcohol measurement” (italics added)]; People v. Lepine (1989) 215 Cal.App.3d 91, 94 [ 263 Cal.Rptr. 543 ] [listing, as additional factors, “the speed of exhalation, the depth of exhalation, the amount of humidity in the air, the amount of mucus in the lungs” (italics added)]; see generally Bransford, supra, 8 Cal.4th at p. 889 ; State v. Brayman (1988) 110 Wn.2d 183 [ 751 P.2d 294, 297 ] (Brayman); Downie, supra, 569 A.2d at pp. 246-248; Annot., supra, 90 A.L.R.4th at p. 160 ; Mason & Dubowski, Breath-Alcohol Analysis, supra, 21 J. Forensic Sciences at pp. 21-29.)

Because the 2,100-to-one standard was set by considering the results from numerous individuals in various performance correlation studies, the general amount of overall variance between the blood-alcohol test results and the breath-alcohol test results as measured by properly working machines became known. (See, e.g., Downie, supra, 569 A.2d at pp. 247-248; Mason & Dubowski, Breath-Alcohol Analysis, supra, 21 J. Forensic Sciences at pp. 16 [table 3], 31-32.) Although most of these performance correlation studies found average partition ratios were around 2,300 to one (e.g., Downie, supra, at p. 247 ), the ratio ultimately chosen by the scientific community for the conversion was set at a lower-than-average level of 2,100 to one. This ratio was established in order to give the benefit of the doubt to the subject in most instances (ibid.; see generally Mason & Dubowski, Breath-Alcohol Analysis, supra, at pp. 23-24), and with the understanding that although for this reason (and various additional reasons) breath-testing results generally underreport actual blood-alcohol concentration, the 2,100-to-one ratio may overestimate actual blood-alcohol concentration in a small percentage of the population— perhaps as low as 0.3 percent of the population according to one analysis, or as high as 2.3 percent of the population in another study. (See Downie, supra, at pp. 247 [describing testimony by Dr. Borkenstein], 248 [describing testimony by Dr. Dubowski].)

We recently reaffirmed the 2,100-to-one conversion ratio, noting that despite the recognized variability, “most scientists” continue to “agree that the . . . ratio roughly approximates or even underestimates the ratio of most people.” (McNeal, supra, 46 Cal.4th at p. 1192.) Similarly, nearly a decade after it decided Downie , the New Jersey Supreme Court recently ordered an evaluation by a special master who in turn heard updated expert opinion concerning this and related issues. Adopting its special master’s findings, the New Jersey court in Chun, supra, 943 A.2d 114 , concluded that the 2,100-to-one ratio remains scientifically valid, and that “[t]he percentage of individuals for whom there may be an overestimation by use of this ratio remains ‘extraordinarily small.’ ” (Id., at p. 139.) The New Jersey high court concluded: “[T]he overwhelming evidence demonstrates that use of this ratio tends to underestimate the actual [blood-alcohol concentration] in the vast majority of persons whose breath is tested. Although, as in Downie , there may be a small number of individuals who are disadvantaged by a device that uses the 2100 to 1 blood/breath ratio, there is sound scientific support for its continued utilization.” (Ibid.; see Brayman, supra, 751 P.2d at pp. 297-303 [rejecting due process and equal protection challenges to use of the 2,100-to-one ratio]; cf. Burg v. Municipal Court (1983) 35 Cal.3d 257 [ 198 Cal.Rptr. 145 , 673 P.2d 732 ] (Burg) [rejecting police power and vagueness or fair notice challenges to the per se statute, and finding it constitutionally permissible to define that offense in terms of a specific alcohol-concentration percentage].)

B. Relevant legislation and related case law

Prior to amendments enacted in 1990, section 23152(b)’s per se offense of driving with a blood-alcohol concentration of 0.08 percent or more—like the statutes in many other states—defined the crime solely in terms of “grams of alcohol per 100 milliliters of blood.” (Stats. 1989, ch. 1114, § 27, p. 4080.) Accordingly, whenever a driver exercised his or her right to elect a breath test over a blood test, it was necessary to “convert the breath results into an equivalent blood-alcohol percentage” (McNeal, supra, 46 Cal.4th at p. 1193, italics added) in order to determine the alcohol concentration in blood by using the standard 2,100-to-one partition ratio. (Id., at p. 1194.) Resulting trials of those charged under such statutory schemes frequently became expensive and time-consuming battles of experts who presented to juries their views concerning the effect of partition ratio variability factors (see ante, pt. II.A.3.) based on the facts of the case. (McNeal, supra, 46 Cal.4th at pp. 1194-1195 [describing this problem under the former Cal. statute]; Bransford, supra, 8 Cal.4th at pp. 888-889 [same].)

1. Amendment of section 23152(b) in 1990 to define the offense alternatively in terms of alcohol concentration in blood or in breath

In response to this and similar litigation occurring nationwide concerning partition ratio variability, Drs. Mason and Dubowski, the authors of the 1976 article cited earlier—Breath-Alcohol Analysis, supra, 21 Journal of Forensic Sciences 9—described the scientific and legal problems triggered by undertaking such conversions and proposed that “[a]ll of these vexing matters [could be] avoided by the expedient of reporting only the quantity of ethanol found per unit volume of delivered breath . . . and defining the offense by statute in terms of the amount of ethanol allowable per unit volume of the sample analyzed.” (Id., at p. 32, italics added.) “Thus for breath, in a given jurisdiction, the quantity of ethanol present in 210 litres of substantially alveolar air (. . . 0.08 g/210 1 for states using the 0.08% [weight/volume] standard for blood), could define by statute the alcohol-related element of the offense of driving while under the influence of alcohol.” (Id., at p. 33.) The authors concluded: “We believe that the conversion of a breath quantity to a blood concentration of ethanol, for forensic purposes, should be abandoned and that the offense of driving while under the influence of alcohol should be statutorily defined in terms of the concentration of ethanol found in the breath in jurisdictions employing breath analysis.” (Ibid., italics added.)

In early 1990, Assembly Bill No. 4318 (1989-1990 Reg. Sess.) was introduced, proposing to revise section 23512(b) by defining the offense of driving while having a prohibited alcohol concentration in terms of “grams of alcohol per 100 milliliters of blood or grams of alcohol per 210 liters of breath.” (As introduced Mar. 2, 1990, § 1, p. 2, amending language in italics.) A report prepared by the Assembly Committee on Public Safety observed that existing California law defined the offense “in terms of the amount of alcohol in the defendant’s blood” and that “[a]ttacks on the partition ratio may result in expensive and time consuming evidentiary hearings and undermine successful enforcement of driving under the influence laws.” (Assem. Com. on Public Safety, Rep. on Assem. Bill No. 4318 (1989-1990 Reg. Sess.) as amended May 15, 1990, p. 2 (Assembly Committee Report).) The report noted that the bill would “[eliminate the need for conversion of a breath quantity to a blood concentration of alcohol by statutorily defining driving under the influence of alcohol in terms of the concentration of alcohol found in the breath when breath analysis is used” (id., at p. 1, italics added)—and in support the report set out, under the heading “Scientific Recommendation,” the above quoted recommendation by Drs. Mason and Dubowski in Breath-Alcohol Analysis, supra, 21 Journal of Forensic Sciences at page 33. (Assem. Com. Rep., supra, at p. 2.) The statute was amended as proposed. (See ante, fn. 2.)

2. Construction of the amended statute in Bransford and McNeal

We construed the revised version of section 23152(b)’s per se offense in Bransford, supra, 8 Cal.4th 885 . We reviewed the history of the bill and concluded that the Legislature intended to criminalize the act of driving with either the specified blood-alcohol level or the specified breath-alcohol level. (Id., at pp. 888-891.) In reaching this determination we acknowledged the defendant’s argument that because the statute on its face continues to “define the offense in terms of blood-alcohol concentration,” it was possible to read the amended statute as merely providing “an alternate means for calculating the blood-alcohol concentration.” (Id., at p. 890, some italics added.) But we concluded that it would be unreasonable to so construe the statute, because the legislative history disclosed an intent “to criminalize the act of driving either with the specified blood-alcohol level or with the specified breath-alcohol level” (ibid., italics added) and hence, we held, the amended statute “prohibited the act of driving with 0.08 percent or more of blood alcohol as defined by grams of alcohol in 210 liters of breath.” (Ibid., original italics.)

Having determined that the amended statute alternatively “defined the substantive offense of driving with a specified concentration of alcohol in the body” ( Bransford, supra, 8 Cal.4th at pp. 892-893, italics added), we also concluded that the amended statute rendered irrelevant consideration of matters such as partition ratio variability, because the revised statute “defined the offense without regard to such ratios.” (Id., at p. 893.) It followed, we held, that expert evidence concerning partition ratio variability was properly excluded in trials under the amended per se statute. (Ibid.)

By contrast, as we subsequently observed in McNeal, supra, 46 Cal.4th 1183 , the traditional generic offense set out in section 23152(a) is defined as driving “while under the influence of alcohol”—and that separate offense is not defined by reference to a prohibited alcohol concentration level. (McNeal, supra, at pp. 1192-1193.) We noted that the subdivision (a) generic offense is further distinguished from the subdivision (b) per se offense of driving with a prohibited concentration of alcohol in that the generic offense carries a rebuttable presumption allowing the jury to presume the defendant is “under the influence” if the jury finds the defendant has a blood-alcohol level of 0.08 percent or more. (McNeal, supra, at pp. 1197-1199 ; see § 23610, subd. (a)(3) [presumption].) We found it significant that this presumption is defined in terms of a blood-alcohol level only—and it is not alternatively defined in terms of a breath-alcohol level. (McNeal, supra, at pp. 1197-1198 .) We observed that although the same conversion factor of 2,100 to one (the standard partition ratio) can be used to translate a breath-alcohol result into a blood-alcohol level, that conversion factor is not a part of the Legislature’s definition of the presumption applicable to the generic offense. (Ibid.) Accordingly, we held in McNeal that whereas evidence of partition ratio variability remains irrelevant and inadmissible with regard to a per se charge of driving with a prohibited concentration of alcohol under section 23152(b), that same evidence is relevant and admissible to rebut the presumption underlying a generic charge of driving under the influence under section 23152(a) when the prosecution relies on the results of a breath-testing machine test. (McNeal, supra, at pp. 1196-1202 ; cf. Bransford, supra, 8 Cal.4th at p. 885 .)

III. Analysis

As noted earlier, the People assert that the appellate court erred both in overturning the trial court’s exclusion of the prehearing and proffered testimony of Dr. Hlastala, and also in finding the alleged errors to be prejudicial. Defendant, on the other hand, insists that neither Dr. Hlastala’s prehearing jury testimony nor the subsequent proposed testimony concerning reliability of breath-testing machines, constituted evidence of partition ratio variability, and that the trial court erred by concluding otherwise and precluding that testimony with regard to the per se section 23152(b) charge. His claims are usefully divided into two categories. First, as mentioned earlier, defendant asserts the trial court’s rulings precluded him from presenting as a defense the argument that all breath-testing machines are unreliable in light of alcohol that is “picked up” during inhalation while traveling past mucus membranes lining the airway, “saturating” the breath before it reaches the alveolar, deep lung regions. Based on this, defendant also planned to argue that these machines fail to collect and test alcohol from alveolar, deep lung air, as required by the first sentence of California Code of Regulations, title 17, section 1219.3, which, as noted earlier, provides that a sample “shall be expired breath which is essentially alveolar in composition.”

Second, defendant asserts, the trial court erred in barring Dr. Hlastala’s additional proposed testimony, adduced at the hearing outside the presence of the jury, that other physiological factors, such as pattern of breathing (speed and depth of inhalation and exhalation), body and breath temperature, and hematocrit level (ratio of red blood cells to total blood volume), make breath testing unreliable.

We will address these contentions in turn. Before doing so, however, we again note that defendant’s trial was conducted before our decision in McNeal, supra, 46 Cal.4th 1183 (discussed ante, pt. H.B.2.), in which we clarified that partition ratio variability evidence, although inadmissible with respect to a per se charge under section 23152(b), is admissible with respect to a generic charge under section 23152(a). Because there was no conviction on the generic charge, we have no occasion to address, with respect to that charge, the propriety of the trial court’s exclusion of the prehearing jury testimony or the other testimony subsequently proffered in this case.

A. Testimony that breath-testing machines are unreliable because expired breath contains only alcohol that has been absorbed from the upper airways and hence the machines fail to sample and analyze the concentration of alcohol contained in alveolar, deep lung air

As observed ante, in part I.B.1., Dr. Hlastala testified before the jury that “little blood vessels . . . called ‘bronchial vessels’ . . . bring alcohol” to the mucus membranes lining the airway, and that “if we inhale and we pick up alcohol from this mucus,” then “by the time we get down to this [alveolar] air sac\\, it’s already filled up and saturated.” (Italics added.) In other words, defendant argues, Dr. Hlastala told the jury (before he was cut off by the court in response to the prosecutor’s objection), that breath-testing machines, because of saturation of breath on inhalation, sample and measure no alcohol from alveolar, deep lung air. Although, as noted earlier, the trial court ultimately did not instruct the jury to ignore Dr. Hlastala’s testimony in this regard, the court did bar any further testimony of this kind. Moreover, as directed by the court, counsel did not argue these aspects of Dr. Hlastala’s testimony before the jury during closing argument. We conclude that, with respect to the statutory per se charge, the trial court did not err in excluding further testimony on this subject or in limiting counsel’s closing arguments to the jury.

Defendant and amicus curiae on his behalf, the California DUI Lawyers Association, argue that this aspect of the excluded expert testimony did not constitute inadmissible partition ratio variability evidence under Bransford, supra, 8 Cal.4th 885 , because in this part of his testimony Dr. Hlastala did not purport to compare breath-testing results with blood-testing results, but instead simply focused on the breath sample that was collected for analysis. In addition, they assert, the evidence was relevant and necessary to allow presentation, as a defense to the per se charge, of the argument that the specimens collected by breath-testing machines fail to comply with California Code of Regulations, title 17, section 1219.3’s specification that breath samples shall consist of essentially alveolar, deep lung breath. Defendant notes that the Legislature (in Health & Saf. Code, § 100700) has mandated compliance with this and various other alcohol-testing regulations adopted by the former State Department of Health Services (now State Department of Public Health), and he relies on People v. Williams (2002) 28 Cal.4th 408, 414, 415-416 [ 121 Cal.Rptr.2d 854 , 49 P.3d 203 ], for the proposition that noncompliance with regulations concerning breath testing itself constitutes relevant evidence that a defendant “may put before the jury because scientific standards behind accuracy are premised on the regulations embodied in title 17.” (See People v. Adams (1976) 59 Cal.App.3d 559, 567 [ 131 Cal.Rptr. 190 ] [when regulations are not followed, defendants are “entitled to attempt to discredit the results by showing that noncompliance affected their validity”].)

The People respond by emphasizing the practical impossibility of sampling pure alveolar, deep lung air: “It would be too invasive to insert a tube into a person’s lungs to attempt to extract pure alveolar air.” (See 2 Defense of Drunk Driving Cases, supra, § 18.01[2], p. 18-7.) They argue that breath-alcohol sampling, using properly calibrated and operated machines such as those at issue here, has long been endorsed by the Legislature, and that these samples have always begun “with an inhalation of outside air which travels down the airway to the deep lungs and back up the airway and out of the body through the mouth. Obviously that airway has always had blood vessels next to it,” along with mucus membranes, “and any effect from such blood vessels occurs in all breath testing.” (Italics added.)

The People elaborate on their position by emphasizing that the underlying performance correlation studies (described ante, pt. II.A.3.) have always used simultaneous (1) blood samples and (2) “end-expiratory” breath samples as measured by properly functioning and calibrated breath-testing machines to verify the accuracy of breath-testing machine readings. The People conclude that “the laws and regulations governing drunk driving and breath alcohol testing do not require the alcohol in breath testing samples to be only from alcohol that originates from the blood vessels next to the alveoli, as opposed to alcohol that originates from the bronchial vessels (bronchial alcohol) and migrates down the airway and into the alveolar region during inhalation. The per se . . . charge of . . . section 23152(b) [merely] prohibits a person from driving while having a breath alcohol level of 0.08 percent or more per 210 liters of breath.”

Defendant and the supporting amicus curiae, in turn, counter by asserting that defendant had a right to challenge the reliability of breath-testing machines generally, and that the People in essence propose to “ignore any problems with the failure of the breath test used in this case to comply with section 1219.3” of California Code of Regulations, title 17.

As explained below, we agree with defendant that this aspect of the prehearing jury testimony was not the same as the type of testimony regarding partition ratio variability that was at issue in Bransford, supra, 8 Cal.4th 885 , and McNeal, supra, 46 Cal.4th 1193 . Nonetheless, we conclude that the expert testimony at issue here was properly excluded, insofar as the trial court’s ruling related to the statutory per se charge.

As noted earlier, we explained in Bransford, supra, 8 Cal.4th 885 , that the 1990 amendment of the per se offense (§ 23152(b)) was specifically designed to obviate the need for conversion of breath results into blood results—and it rendered irrelevant and inadmissible defense expert testimony regarding partition ratio variability among different individuals or at different times for the same individual. ( Bransford, supra, at p. 893 .) The revised statute defines the 0.08 percent breath-alcohol offense in light of the standard partition ratio, which in turn gives the benefit of the doubt to defendants and understates the equivalent blood-alcohol concentration for all but a very small percentage of the population. As we observed in Burg, supra, 35 Cal.3d 257 , defining the per se offense in terms of a specific alcohol-concentration percentage complies with constitutional requirements concerning vagueness and fair notice. (See ante, fn. 17.)

Although the prehearing jury testimony trader discussion at this point did not purport to address the variability of partition ratios for different persons or at different times for the same individual, the testimony was nonetheless analogously and fundamentally at odds with the statutory per se offense, in that it suggested the results of breath-testing machines always are unreliable because they fail to measure the alcohol content of the air from the alveolar region of the lungs. That testimony failed to take into account that the standard 2,100-to-one partition ratio expressly incorporated in the per se statute was established by simultaneously measuring a subject’s blood-alcohol level and breath-alcohol level as disclosed by a properly working and calibrated breath-testing machine that samples the last part of the subject’s expired breath. Accordingly, whether or not that part of expired breath accurately reflects the alcohol that is present only in the alveolar region of the lungs, the statutorily proscribed amount of alcohol in expired breath corresponds to the statutorily proscribed amount of alcohol in blood, as established by the per se statute.

Applying standard principles of statutory construction to determine and effectuate legislative intent (e.g., People v. Pieters (1991) 52 Cal.3d 894, 898-899 [ 276 Cal.Rptr. 918 , 802 P.2d 420 ]), we conclude that when the Legislature employed the word “breath” in section 23152(b), it had in mind the air that is exhaled into a properly working and calibrated breath-testing machine. It follows that, like the expert testimony regarding partition ratio variability that was held irrelevant and inadmissible in Bransford, the expert testimony under review here is similarly inadmissible insofar as it relates to the statutory per se charge because it as well conflicts with the underlying premise and definition of the per se offense.

As noted earlier, defendant also observes that the Legislature, in Health and Safety Code section 100700 (quoted ante, fn. 22) has mandated compliance with the various alcohol-testing regulations adopted by the State Department of Health Services. Based on this, defendant additionally claims that the expert evidence should be admissible, concerning the statutory per se charge, to show that breath-testing machine results are inconsistent with the governing state regulations—especially California Code of Regulations, title 17, section 1219.3, which as observed calls for a sample of “essentially alveolar” breath. As we will explain, defendant’s contention rests on an incorrect understanding and interpretation of the regulations in question.

Initially, we observe that although many other jurisdictions employ the machines used here and similar models, neither defendant nor his amicus curiae cites any decision from any jurisdiction allowing testimony like that offered here, challenging generally the reliability of breath-testing machines that have been properly maintained and operated, on the basis that the alcohol content of the breath measured by the machines assertedly is only that which is absorbed from the upper airways and not the alveolar region. More specifically, although most jurisdictions have regulations essentially identical to California Code of Regulations, title 17, section 1219.3 (see ante, fns. 9-11), neither defendant nor his amicus curiae cites any decision from any jurisdiction allowing testimony like that at issue here, telling a jury that properly functioning and operated breath-testing machines sample no alcohol from alveolar, deep lung air. Indeed, most scientists and courts have long concluded otherwise. (See BASS for Qualification Testing of Breath Alcohol Measurement Devices, supra, p. vii [properly designed breath-testing machines analyze “a sample that is essentially ‘deep lung’ or alveolar air”]; McNeal, supra, 46 Cal.4th at p. 1188 [a breath-testing machine “measures the amount of alcohol vapor expelled into alveolar spaces deep in the lungs”]; Brayman, supra, 751 P.2d at p. 297 [breath-testing machines “are designed to test the last portion of a person’s breath . . . , which comes from deeper portions of the lungs, the alveolar sacs, where the alcohol is transferred from the blood to the lung air”]; Chun, supra, 943 A.2d at p. 127 [the samples collected by breath-testing machines focus on the last part of an exhalation, which “comes from deeper in the respiratory system, where it contains alcohol that more closely represents the amount passing through the lungs from the circulating blood”].)

We also find it significant that for many decades it has been understood that the passage of breath through the airways does indeed influence the resulting sample and makes it “virtually impossible ... to obtain” samples of air from the alveolar region that are “uncontaminated” by the exposure to mucus membranes lining the airways because “[t]he sample taken at the end of the expiration must necessarily pass through the bronchi and oral-pharyngeal cavities and in doing so may gain or lose gases to the gas-fluid interfaces of the mucous membranes.” (Spector, Alcohol Breath Tests: Gross Errors in Current Methods of Measuring Alveolar Gas Concentrations (Apr. 2, 1971) 172 Science 57, 59; see Dubowski, The Technology of Breath-Alcohol Analysis, supra, U.S. Dept, of Health & Human Services, pub. No. (ADM)92-1728, at p. 5 [“alveolar air is modified ... by repeated reequilibration of the alcohol of the breath with that of the fluids lining the respiratory tract during expiration . . .”].)

These considerations demonstrate that when title 17, section 1219.3 of the California Code of Regulations was adopted in the early 1970s, it was contemplated that a breath sample would be obtained by a properly functioning and calibrated breath-testing machine such as those at issue here—devices that have always operated on the fundamental principle of recovering the last part of an expired breath that has traveled from the alveolar region and then out through the upper airway. The fact that the regulation speaks of a breath sample consisting of “expired breath which is essentially alveolar in composition” (italics added) underscores that the regulation contemplates a breath sample consisting of the last part of breath (specifically, the part of exhaled breath measured by a properly operating breath-testing machine), which the regulation identified and described as “essentially alveolar in composition.” For these reasons, and again employing standard principles of construction, we conclude the regulation’s command for such samples must properly be understood as calling simply for a specimen consisting of the last portion of expired breath that is captured by a properly working and calibrated breath-testing machine.

Given this understanding of California Code of Regulations, title 17, section 1219.3, we reject defendant’s argument that the proffered expert testimony was relevant and admissible with regard to the statutory per se charge to demonstrate that the breath-testing machine results were obtained in violation of the regulation in question. Indeed, the prehearing jury testimony appears to stand in direct conflict with the legislatively endorsed regulation, inasmuch as the prehearing jury testimony asserted that no part of expired breath is “essentially alveolar in composition”—which would mean that no breath sample would ever satisfy the regulation’s requirement. “We presume the promulgating body did not intend that result.” (State v. Esser (Ct.App. 2003) 205 Ariz. 320 [ 70 P.3d 449, 452 ] [rejecting Dr. Hlastala’s proposed testimony in that case as inconsistent with the analogous Ariz. breath-sample regulation].)

Further, in the process of considering this evidentiary question, our independent research revealed other regulations and statutes, not addressed in the parties’ original briefs, that shed additional light on how the Legislature views the scheme of breath testing for alcohol concentration. These regulations and statutes confirm our interpretation of title 17, section 1219.3 of the California Code of Regulations, and our conclusion that the trial court did not err in this aspect of its evidentiary mling.

1. Regulations amended in 1985: California Code of Regulations, title 17, sections 1221.2 and 1221.3

Through the 1970s, the State Department of Health Services, acting under then applicable legislative authorization (see Health & Saf. Code, former §§ 436.52, 436.50), adopted its own breath-testing standards and maintained its own list of approved breath-testing machine models. (See Cal. Code Regs., tit. 17, former §§ 1221.2 [“Standard of Performance”], 1221.3 [“Instrument Evaluations”], both set out as amended in Register 75, No. 48 (Nov. 29, 1975) pp. 88.4.13-88.4.15.) In the early 1980s, the State Department of Health Services began considering the idea of deferring to the evolving breath-testing standards and list of approved machine models established by the federal National Highway Traffic Safety Administration. (See, e.g., notes of meetings of Cal. State Dept, of Health Services, Advisory Com. on Alcohol Determination (Dec. 14, 1982), Cal. State Dept, of Health Services, Advisory Com. on Alcohol Determination, Ad Hoc Com. on Breath Testing (Aug. 31, 1982).) By the mid-1980s an emergency situation existed: The State Department of Health Services had not approved new machine models for use in California since 1979, and law enforcement agencies throughout the state were running out of working approved devices. Accordingly, the department amended and adopted California Code of Regulations, title 17, sections 1221.2 and 1221.3, as an emergency measure in late 1985. (See, e.g., Cal. Reg. Notice Register 86, No. 1-Z (Jan. 3, 1986) B-14 & B-15 [emergency order].)

As amended in 1985, and. as still in effect today, California Code of Regulations, title 17, section 1221.2 reads: “(a) Instruments for breath alcohol analysis shall meet the following standard: [f] (1) The instrument and any related accessories shall be capable of conforming to the ‘Model Specifications for Evidential Breath Testing Devices’ of the National Highway Traffic Safety Administration of the U.S. Department of Transportation, which were published in the Federal Register, Vol. 49, No. 242, Pages 48854-48872, December 14, 1984, and are hereby adopted and incorporated, [f] (b) The ability of instruments and any related accessories to conform to the standard of performance set forth in this section shall be tested by the U.S. Department of Transportation.” (See Register 85, No. 52 (Dec. 28, 1985) p. 88.4.13 [eff. on filing], italics added.) At the time of trial in this case, and today, the cited federal regulation provided that breath-testing machines must “measure the alcohol content of deep lung breath samples with sufficient accuracy for evidential purposes.” (58 Fed.Reg. 48705, 48707 (Sept. 17, 1993) [Model Specifications for Evidential Breath Testers], italics added.)

Correspondingly, as amended in 1985, and still today, California Code of Regulations, title 17, section 1221.3 reads: “(a) Only such types and models of instruments and related accessories as are named in the ‘Conforming Products List’ published in the Federal Register by the National Highway Traffic . . . Safety Administration of the U.S. Department of Transportation shall be used for breath alcohol analysis in this State.” (See Register 85, No. 52 (Dec. 28, 1985) p. 88.4.14 [eff. on filing].)

The notice accompanying the State Department of Health Services’s emergency order explained: “The instruments on the [federal] list have passed laboratory evaluation by DOT [(the U.S. Dept, of Transportation)] against [that agency’s] performance specifications .... The use of these instruments will provide law enforcement agencies in the State a list of current, state-of-the-art approved breath testing instruments. [][] The Department and DOT approval mechanisms for the breath testing instruments differ significantly only in the experimental test for establishing the correlation of the instruments’ breath test results with the actual alcohol concentrations in the blood of the human test subjects. The Department uses tests of human subjects for this determination while DOT uses a Breath Alcohol Sample Simulator (BASS). The BASS is a device designed by DOT to provide the appropriate breath sample by simulating the variables of a human expiration including pressure, temperature, and rate of delivery. The Advisory Committee on Alcohol Determination has determined that the instruments approved by either test are equivalent in terms of scientific validity.” (Cal. Reg. Notice Register 86, No. 1-Z (Jan. 3, 1986), A-32 [Informative Digest], italics added; see generally BASS for Qualification Testing of Breath Alcohol Measurement Devices, supra, at p. vii [noting that the BASS device is designed to “provide an objective, reproducible mechanical substitute for human subjects in the evaluation of the performance of the breath sample collecting systems of breath alcohol instruments”].)

2. Legislative action in 2004: Health and Safety Code section 100701

Nearly two decades after the State Department of Health Services adopted these emergency regulations, the Legislature enacted Health and Safety Code section 100701 in 2004. (Stats. 2004, ch. 337, § 4, pp. 3323-3324, enacting Sen. Bill No. 1632 (2003-2004 Reg. Sess.) [Sen. Johnson, bill author].) The statute requires that all “breath alcohol instruments and calibrating devices used in testing” be those set out in the “conforming products list in the Federal Register by the National Highway Traffic Safety Administration of the United States Department of Transportation.” (Health & Saf. Code, § 100701.)

As observed earlier, another statute, Health and Safety Code section 100700 (quoted ante, fn. 22), also adopted in the same 2004 legislation, reiterates the long-standing policy of mandating compliance with the various alcohol-testing regulations adopted by the State Department of Health Services as set out in California Code of Regulations, title 17, “commencing with Section 1215”—the group of regulations applicable here. Among those regulations are not only the provision upon which defendant focuses—the first sentence of section 1219.3—but also sections 1221.2 and 1221.3, which, as discussed above, adopt the corresponding federal regulations governing, respectively, specifications for breath-testing machines and approved models for use.

In other words, for almost two decades before the Legislature enacted Health and Safety Code sections 100700 and 100701 in 2004, the State Department of Health Services had itself adopted the applicable federal specifications for breath-testing machines, and the approved model list of those machines. Accordingly, when the Legislature acted in 2004, it in essence ratified those earlier—and, by that time, well-tested—administrative determinations.

3. The effect of these regulations and statutes

We asked the parties to submit supplemental briefs concerning the effect, if any, of these statutes and related regulations. Defendant concludes that although the People may be expected to argue otherwise, these statutes and regulations do not support an assertion that the trial court properly excluded the challenged testimony insofar as it related to the statutory per se charge. Rather more surprisingly, the People—represented by the City Attorney of San Diego—do not argue otherwise. They assert that these statutes and regulations merely have the “combined effect of ensuring the accurate testing of deep lung, or alveolar, air.” As explained below, we disagree with both parties.

As noted earlier, under the applicable regulations an officer observes a subject for 15 minutes prior to testing in order to ensure that the resulting sample of “end-expiratory” deep lung, alveolar breath is not contaminated by mouth alcohol or regurgitation. Further, machines like the Intoximeter EC/IR used in this case are designed to indicate whether an adequate sample of end-expiratory breath has been provided—and to prompt the operator to secure a subsequent sample as necessary.

It is apparent that the models of breath-testing machines employed here— and scores of others—have been subject to rigorous review and have been found by the federal agency to reliably measure the alcohol content of alveolar, deep lung breath samples. (See generally 58 Fed.Reg. 48705, 48706-48707 [“Model Specifications”—procedures for submitting machines for testing], 48707-48708 [“Model Specifications”—definitions and eight specific tests].) As observed above, no machine model can meet federal standards unless it “measure[s] the alcohol content of deep lung breath samples with sufficient accuracy for evidential purposes.” (58 Fed.Reg. 48705, 48707, italics added.) The circumstance that both machine models used in this case have been certified for evidential use (see ante, fn. 24) reflects that they have been found by the National Highway Traffic Safety Administration of the United States Department of Transportation to reliably sample and measure alveolar, deep lung breath as contemplated by both federal and California regulations.

Even more significantly, it also is apparent that the State of California has enacted a policy of adopting, incorporating, and deferring to federal specifications for, and certification of, models of breath-testing machines. By both statute and legislatively endorsed regulations, California has in essence determined that all models meeting the federal standards produce sufficiently reliable results for purposes of California’s statutes relating to alcohol-concentration limits. To the extent Dr. Hlastala’s testimony would have informed the jury that the breath-testing machines employed in this case, and others, are unreliable because breath becomes saturated with alcohol while being inhaled, and hence the machines fail to measure any alveolar, deep lung air, the proposed testimony sought to nullify determinations necessarily made by the federal agency, adopted by the State Department of Health Services, and endorsed by our Legislature in 2004, when, presumably aware of the science reporting that breath samples are by definition affected by the travel of breath through the respiratory system (see, e.g., Spector, Alcohol Breath Tests: Gross Errors in Current Methods of Measuring Alveolar Gas Concentrations, supra, 172 Science at p. 59), the Legislature nevertheless implicitly endorsed the State Department of Health Services conclusion that any machine that meets the federal agency’s specifications and is listed as a conforming model is reliable and approved for evidential use in a California prosecution.

Although Dr. Hlastala may hold scientifically based reservations concerning these legislative conclusions, we must defer to and honor the Legislature’s reasonable determinations made in the course of its efforts to protect the safety and welfare of the public. As our Court of Appeal has observed, “The fact that the current state of scientific knowledge has not settled the ongoing scientific debate as to the best method of measuring inebriation does not preclude the Legislature from regulating driving based on conflicting scientific theories.” (People v. Ireland (1995) 33 Cal.App.4th 680, 693 [ 39 Cal.Rptr.2d 870 ].) Likewise, as the Washington Supreme Court observed in analogous circumstances: When “scientific opinions conflict on a particular point, the Legislature is free to adopt the opinion it chooses, and the court will not substitute its judgment for that of the Legislature.” ( Brayman, supra, 751 R2d at p. 300; see Downie, supra, 569 A.2d at p. 251 [concluding that “[t]he reliability of [breath-testing machine testing] results will continue to be the subject of judicial notice . . .” and that testimony concerning partition ratio variability by expert witnesses “will. . . continue to be inadmissible” in prosecutions under a “per se” statute].)

As the trial court observed, defendant remained free to argue, and present evidence, that the particular machines used in this case malfunctioned, or that they were improperly calibrated or employed. But the fundamental reliability of the breath-testing machine models used in this case to produce results that are pertinent to the per se statute has been determined by the Legislature. That legislative determination is not subject to rebuttal as a defense in a criminal prosecution. It is hornbook law that a “court’s authority to second-guess the legislative determinations of a legislative body is extremely limited. It is a ‘well-settled principle that the legislative branch is entitled to deference from the courts because of the constitutional separation of powers.’ ” (Connecticut Indemnity Co. v. Superior Court (2000) 23 Cal.4th 807, 814 [ 98 Cal.Rptr.2d 221 , 3 P.3d 868 ]; see, e.g., Minnesota v. Clover Leaf Creamery Co. (1981) 449 US. 456, 469 [ 66 L.Ed.2d 659 , 101 S.Ct. 715 ] [deferring to legislative determinations so long as they are “ ‘at least debatable’ ”].) Just as a court will not substitute its judgment for that of the Legislature in this regard, nor may a defendant, through an expert witness, invite a jury to substitute the expert’s judgment for that of the Legislature in this respect.

For these reasons we conclude that defendant’s regulation-based argument in support of admitting the challenged expert evidence rests on an incorrect understanding and interpretation of the legislatively endorsed regulations in question. It follows that under both the statute (§ 23152(b)) and the regulation on which defendant relies (Cal. Code Regs., tit. 17, § 1219.3) the trial court properly precluded and restricted Dr. Hlastala’s challenged testimony insofar as it related to the statutory per se charge.

B. Proposed testimony that breath-testing machines are unreliable because other physiological factors may affect the transmission of alcohol from the bloodstream to the deep portions of the lungs and then through the exhalation process

As noted earlier, defendant also sought to have Dr. Hlastala testify that a series of other physiological factors—individual breathing patterns (speed and depth of exhalation), body and breath temperature, sex, and hematocrit level (ratio of red blood cells to total blood volume)—may affect the transmission of alcohol from the bloodstream to the deep portions of the lungs and then through the upper airway and finally through the mouth, thereby rendering unreliable all machines that sample and test breath-alcohol specimens.

Although many other jurisdictions employ these and similar breath-testing machines, defendant cites only one decision from another jurisdiction appearing to allow testimony like that offered here—challenging, on the basis of the listed physiological factors, the reliability of approved devices that have been properly maintained and operated. (State v. Cooperman (Ct.App. 2012) 230 Ariz. 245 [ 282 P.3d 446 ] (Cooperman) [affd., without addressing this issue, in State v. Cooperman (2013) 306 P.3d 4 ].)

As explained below, we do not find this authority persuasive with regard to the evidentiary question under California law. (See post, fns. 30 & 31.) Instead, we agree with the trial court that this aspect of the expert testimony would essentially constitute partition ratio variability evidence, which, as noted, is barred in section 23152(b) per se prosecutions under Bransford, supra, 8 Cal.4th 885 .

As the Court of Appeal below emphasized, and as defendant stressed in his briefs and at oral argument, Dr. Hlastala expressly disclaimed that his testimony implicated the concept of partition ratio variability, because, he said, he was not questioning the conversion of breath-alcohol levels into blood-alcohol levels. Instead, he asserted, he was considering factors that influence the composition of the breath sample that the machines collect. In other parts of Dr. Hlastala’s proposed testimony, however, he did make repeated references to comparing breath tests to blood tests—maintaining that (1) breath testing is inferior because it is based on the assumption that the measured breath “is directly related to water in the lungs, which is directly related to what’s in the blood” and (2) as between testing of venous blood and breath testing, the former was “the standard,” and superior. The evident import of these aspects of his proposed testimony was that breath testing is suspect in comparison with venous blood testing, and that in order to be scientifically reliable, a breath test result should not be considered on its own, but only in relation to the known result of a blood test. In this sense, Dr. Hlastala’s proposed testimony would indeed have invited the jury to draw unfavorable comparisons between the two modes of testing that the Legislature has, by section 23152(b), found to be legally equivalent alternatives.

Moreover, although Dr. Hlastala sought to narrowly define the concept of partition ratio variability testimony (as limited to testimony that breath test results vary when compared to blood test results taken from the same person at the same time) in order to avoid the bar concerning such testimony that we confirmed in Bransford, supra, 8 Cal.4th 885 , the prohibition that we articulated in that case should reasonably be understood more broadly, as extending to expert testimony concerning the underlying factors that cause variability in breath-alcohol samples. As observed ante, in part I.B.2., the factors that Dr. Hlastala focused upon in this aspect of his proposed testimony—pattern of breathing (speed and depth of exhalation), body and breath temperature, hematocrit level (ratio of red blood cells to total blood volume), as well as the other factors that he conceded, on questioning, to be relevant to his general point, such as sex, medical condition, such as lung disease—are prominent among the factors that scientists and courts have long recognized as influencing partition ratio variability. (See ante, pt. II.A.3., at p. 18.) These factors affect the results of breath-testing machines generally—and thus have already been taken into account by the widely accepted statutory partition ratio and the performance, or correlation, studies discussed earlier.

The People correctly note that defendant—the proponent of the evidence— had the burden to produce preliminary facts to support his view that this aspect of Dr. Hlastala’s proposed testimony did not constitute partition ratio variability evidence. Dr. Hlastala’s observation that there exists variation in the alcohol concentration of any given breath sample is, in essence, substantively similar to observing that there is variation in partition ratios among persons and indeed in a single person—testimony that is inadmissible as irrelevant in a section 23152(b) per se prosecution under Bransford, supra, 8 Cal.4th 885 . Although Dr. Hlastala attempted to distinguish (1) factors that influence the composition of the breath sample that the machines collect from (2) the problem of partition ratio variability—and asserted that he was speaking only to the first, and not the second—neither defendant nor Dr. Hlastala ultimately offered evidence to establish that the overlapping factors somehow affect variability associated with the composition of breath samples in any way that is separate and distinct from the way those same factors affect variability associated with the standard partition ratio. Under these circumstances, we conclude that the trial court properly equated the proposed testimony with partition ratio variability evidence and properly excluded it with respect to the statutory per se charge—and that the Court of Appeal below erred in concluding otherwise.

In any event, the reasons we have discussed ante, in part III.A., also clearly support a conclusion that the trial court properly excluded the challenged testimony. To the extent that Dr. Hlastala’s testimony would have informed the jury that all results from breath-testing machines are unreliable because of the listed physiological factors, the proposed testimony sought to nullify determinations necessarily made by the federal agency and endorsed by our Legislature in 2004, when, presumably aware of the science and decisions reporting that breath samples are affected by these same and related factors (see ante, pt. II.A.3.), the Legislature nevertheless implicitly endorsed the State Department of Health Services’s 1985 determination that any machine model that meets the federal agency’s specifications and is listed as conforming is reliable and approved for evidential use in a California prosecution. As noted earlier, just as that legislative determination is not subject to nullification by courts, nor is it subject to nullification by a jury at the invitation of an expert witness in a criminal prosecution of a statutory per se charge.

IV. Conclusion and disposition

The trial court did not err in limiting Dr. Hlastala’s prehearing jury testimony and excluding his subsequently proposed elaborating testimony with respect to the statutory per se charge. As the trial court observed, defendant remained free to argue, and present evidence, that the particular machines used in this case malfunctioned, or that they were improperly calibrated or employed. But as explained earlier, the 0.08 percent breath-alcohol concentration formulated by the Legislature in enacting the underlying per se offense, section 23152(b), was adopted on the basis of correlation studies employing just such breath-testing machines—and the various physiological factors that affect the results of breath-testing machines generally, have already been taken into account by those studies and the widely accepted statutory partition ratio. We construe both the statute, section 23152(b), and the regulation on which defendant relies, California Code of Regulations, title 17, section 1219.3, as calling simply for a breath specimen consisting of the last portion of expired breath that is captured by an approved breath-testing machine that is properly calibrated and employed.

In light of these conclusions and the corresponding regulations and statutes discussed earlier, the fundamental reliability of federally approved, properly calibrated and employed breath-testing machines used in the application and enforcement of the per se statute is a matter that has been determined as policy by the Legislature—and a defendant’s expert witness may not invite a jury to nullify that determination in the manner at issue here. Accordingly, the judgment of the Court of Appeal is reversed.

Kennard, J., Baxter, J., Werdegar, J., Chin, J., Corrigan, J., and Liu, J., concurred.

See Vehicle Code section 23612, subdivision (h) (a preliminary alcohol screening test (PAS) is an investigative tool used to determine whether there is reasonable cause for arrest). As explained in 72 Ops.Cal.Atty.Gen. 226 , 227 (1989), a preliminary test is “distinguished from the chemical testing of a driver’s blood, breath or urine contemplated by the implied consent law (Veh. Code, § [23612]) which is administered after the driver is arrested, [and is] sometimes referred to as ‘evidentiary’ [or evidential] testing.” (Underscoring omitted.) Section 23152(a) reads: “It is unlawful for any person who is under the influence of any alcoholic beverage or drug, or under the combined influence of any alcoholic beverage and drug, to drive a vehicle.” Section 23152(b) reads: “It is unlawful for any person who has 0.08 percent or more, by weight, of alcohol in his or her blood to drive a vehicle. ['][] For purposes of this article . . . , percent, by weight, of alcohol in a person’s blood is based upon grams of alcohol per 100 milliliters of blood or grams of alcohol per 210 liters of breath, [f] In any prosecution under this subdivision, it is a rebuttable presumption that the person had 0.08 percent or more, by weight, of alcohol in his or her blood at the time of driving the vehicle if the person had 0.08 percent or more, by weight, of alcohol in his or her blood at the time of the performance of a chemical test within three hours after the driving.” The amended complaint also charged one count of driving in excess of 100 miles per hour (§ 22348, subd. (b)). Defendant’s motion argued that evidence of partition ratio variability should be excluded only in the context of the per se offense, and sought an instruction limiting such evidence to the generic charge. We observe that trial was conducted before our decision in People v. McNeal (2009) 46 Cal.4th 1183 [ 96 Cal.Rptr.3d 261 , 210 P.3d 420 ] (McNeal), in which we clarified that partition ratio variability evidence, although (1) inadmissible with respect to a per se offense of driving with 0.08 percent or more blood-alcohol concentration under section 23152(b) as determined in People v. Bransford (1994) 8 Cal.4th 885 [ 35 Cal.Rptr.2d 613 , 884 P.2d 70 ] (Bransford), (2) is admissible with respect to a generic offense of driving under the influence under section 23152(a). Both McNeal and Bransford are discussed post, in part II.B.2. Detailed regulations concerning maintenance and calibration of breath-testing machines are set out in California Code of Regulations, title 17, section 1221.4, which lists, among various standards of procedure, requirements including (1) “accuracy and precision limits of plus or minus 0.01 grams % of the true value” and (2) periodic testing and calibration of machines, performed within every 10 days or 150 subject-testing uses by qualified and trained operators. The concepts of “partition ratio,” and “partition ratio variability,” are discussed post, in part II.A.3. Our decision in Bransford is discussed post, in part II.B.2. Ultimately, the court did not instruct the jury to disregard any of Dr. Hlastala’s testimony. And yet, as directed by the court, counsel for the parties confined their closing arguments to the jury—insofar as they concerned Dr. Hlastala’s opinions—to his unobjected-to testimony (described post, pt. I.B.3.) about the possible effects of mouth-alcohol sample contamination, and related matters. California Code of Regulations, title 17, section 1220.4 governs “expression of analytical results.” It provides in subdivision (a) that blood-alcohol analysis is expressed “in terms of the alcohol concentration in blood, based on the number of grams of alcohol per 100 milliliters of blood.” The regulation provides in subdivision (f): “[B]reath alcohol concentration shall be converted to an equivalent blood alcohol concentration by a calculation based on the relationship: the amount of alcohol in 2,100 milliliters of alveolar breath is equivalent to the amount of alcohol in 1 milliliter of blood.” The regulation also provides in subdivision (e): “[Ujrine alcohol concentration shall be converted to an equivalent blood alcohol concentration by a calculation based on the relationship: the amount of alcohol in 1.3 milliliters of blood is equivalent to the amount of alcohol in 1 milliliter of urine.” The first two of these three provisions are based on the governing statute, Vehicle Code section 23610, subdivision (b), which provides: “Percent, by weight, of alcohol in [a] person’s blood shall be based upon grams of alcohol per 100 milliliters of blood or grams of alcohol per 210 liters of breath.” See, e.g., 58 Federal Register 48705, 48707 (Sept. 17, 1993) (“Model Specifications for Devices to Measure Breath Alcohol” promulgated by the Nat. Highway Traffic Safety Admin., U.S. Dept, of Transportation: breath-testing machines must “measure the alcohol content of deep lung breath samples with sufficient accuracy for evidential purposes”); 5 Code of Colorado Regulations section 1005-2, part 4.3.1 (calling for “delivery of a breath sample that contains end-expiratory air from the lungs”); Administrative Rules of Montana, rules 23.4.201(9) (calling for devices “capable of capturing and analyzing deep lung air to establish the concentration of alcohol contained in that sample”), 23.4.201(15) (“ ‘Deep lung air’ means air which comes from the deeper section of the lung and contains only a portion of alveolar air. This is the type of breath captured by the breath analysis instrument.”); New Mexico Administrative Code section 7.33.2.15 B(2) (sample is to be “end expiratory in composition”). See Arizona Administrative Code section R13-10-103.B1 (“alveolar in composition”); Oklahoma Administrative Code section 40:30-l-3(c)(2) (“consisting substantially of expired alveolar air”); West Virginia Code of State Rules section 64-10-7.1(a) (“specimens which approach alveolar composition”). Other state regulations specify that “Breath means exhaled lung air that. .. contains a large portion of air from the alveolar region of the lungs where the exchange of gases between the blood and air occurs.” (Neb. State Admin. Rules & Regs., tit. 177, ch. 1, § 001.04.) See 501 Code Massachusetts Regulations section 2.05 (device shall be capable of analyzing “samples of alveolar or deep lung[] air”); Ohio Administrative Code section 3701-53-02(C) (“deep lung (alveolar) air”); Wyoming Rules and Regulations, Department of Health, chapter 2, section 2(a) (“deep lung breath (substantially alveolar in composition)” [chemical analysis of blood-alcohol program]); Utah Administrative Code R714-500-6.A.3, R714-500-7.B. (“essentially alveolar or end expiratory in composition”). See 007 21 Code of Arkansas Rules and Regulations 001, 4.12a.(l) (breath sample must be “essentially alveolar in composition” [blood-alcohol program]); Michigan Administrative Code, R 325.2651, rule l.(l)(j) (“essentially alveolar breath”); 37 Texas Administrative Code section 19.2(b)(1) (“essentially alveolar or deep lung air in composition”). Still other state regulations simply provide that a breath sample must meet the requirements of an approved breath-testing machine. (See, e.g., 13 Alaska Admin. Code, § 63.040 [“satisfactory sample” is obtained when the machine’s “visual display” so indicates]; Fla. Admin. Code R. 11D-8.002(12) [“samples of breath . . . using an approved breath test instrument”]; Md. Code Regs. 10 .35.02.02.B.(5) [“ ‘Breath sample’ means the amount of breath delivered by the individual being tested that is an adequate volume to be analyzed by the breath testing instrument.”].) See generally Mason and Dubowski, Breath-Alcohol Analysis, supra, 21 Journal of Forensic Sciences at pages 23-24 and table 4 on pages 26-27 (describing the history underlying the general scientific consensus supporting the 2,100-to-one ratio). In light of the nature of this performance correlation methodology, an awareness of every factor affecting the variability and the particular contribution each factor made to the variability was neither known nor required. The goal of these studies was simply to arrive at an empirically determined scientific formula for a breath test result that would correspond in a reasonably acceptable manner to a venous blood test result. As explained in Downie, supra, 569 A.2d at page 247 , other factors causing breath-testing machines to render “many more results on the low side than on the high side” include (1) the fact that breath-testing machine “results are truncated, or the third decimal position is dropped when read. If a person reads .099 on the [breath-testing machine], the results will be shortened to read .09, thereby underestimating the breath alcohol” and (2) “a suspect may not provide enough deep breath to register all of the alcohol present in the alveolar air. . . .” (Ibid.) And of course, under the implied consent statute, any person who believes that he or she falls into the small minority of persons who might be disadvantaged by a breath test remains free to decline to take such a test, and insist on a blood test only. (See § 23612.) In rejecting the equal protection challenge, the court observed: “The record contains substantial evidence linking driving impairment with the statutorily prohibited breath alcohol level. Although the record contains evidence that breath alcohol may not accurately show blood-alcohol in particular individuals, the record does not establish that proscribing a particular level of breath alcohol is wholly irrelevant to achieving the purpose of reducing DWI hazards.” (Brayman, supra, 751 P.2d at p. 303 .) In rejecting the vagueness or fair notice challenge to the per se statute in Burg , we observed: “One who drives a vehicle after having ingested sufficient alcohol to approach or exceed the level proscribed is neither ‘innocent’. . . nor is he without ‘fair warning.’ ... It is difficult to sympathize with an ‘unsuspecting’ defendant who did not know if he could take a last sip without crossing the line, but who decided to do so anyway, [f] The very fact that he has consumed a quantity of alcohol should notify a person of ordinary intelligence that he is in jeopardy of violating the statute. . . . Considering also today’s heightened level of public awareness regarding the problem, we cannot believe that any person who drives after drinking would be unaware of the possibility that his blood-[ or breath-]alcohol level might equal or exceed the statutory standard.” (Burg, supra, 35 Cal.3d at p. 271 , fns. & citations omitted.) Similarly, in the present case, although the 2,100-to-one standard ratio embodied in the per se statute may overstate the breath- and blood-alcohol level for a relatively few individuals, even those individuals have put others at risk by driving after drinking. Additional relevant legislation and regulations not addressed by the parties in their original briefs will be discussed post, part III.A. We observed in closing that the “defendants remained free to challenge the breath-test results on other, relevant grounds, including the reliability of the machine and the manner in which the test was administered.” (Bransford, supra, 8 Cal.4th at p. 893 , italics added.) By the italicized phrase, we contemplated that a defendant may challenge whether the particular machine actually employed to collect and analyze his or her breath sample was unreliable because it was not calibrated or maintained consistently with applicable standards and regulations. (E.g., Cal. Code Regs., tit. 17, § 1221.4; see ante, fn. 4.) In other words, for purposes of rebutting the presumption, a defendant has a right to argue and prove that he or she is among the small percentage of people (see ante, pt. II.A.3.) who might suffer an overstated result by virtue of the standard 2,100-to-one partition ratio. “If the defendant’s own partition ratio is the same as or higher than this standard ratio, his converted blood-alcohol result is valid or skewed in his favor. If, however, the defendant’s own partition ratio at the time of testing is lower than the standard ratio, conversion of the breath result using the statutory formula produces an artificially high measure of his blood alcohol. []Q . . . . For someone with an extremely low ratio of 1,100 to 1, for example, use of the 2,100-to-l partition ratio would overstate blood-alcohol content by almost 50 percent.” (McNeal, supra, 46 Cal.4th at p. 1198 .) Moreover, we noted: “Evidence showing the defendant had a low partition ratio, and thus a lower concentration of blood alcohol than was reported, could also support an inference that he was not under the influence in violation of the generic DUI law.” (Id.., at pp. 1198-1199.) We acknowledged that, in light of this and evidence of variability of partition ratios in the general population, “there is a possibility that the blood-alcohol level such a machine produces for a given defendant is inaccurate.” (Id., at p. 1199.) The regulation addresses three discrete aspects of “[b]reath [sample] [collection,” and reads in full: “A breath sample shall be expired breath which is essentially alveolar in composition. The quantity of the breath sample shall be established by direct volumetric measurement. The breath sample shall be collected only after the subject has been under continuous observation for at least fifteen minutes prior to collection of the breath sample, during which time the subject must not have ingested alcoholic beverages or other fluids, regurgitated, vomited, eaten, or smoked.” (Cal. Code Regs., tit. 17, § 1219.3.) The statute provides in relevant part: “Laboratories engaged in the performance of forensic alcohol analysis tests by or for law enforcement agencies on blood, urine, tissue, or breath for the purposes of determining the concentration of ethyl alcohol in persons involved in traffic accidents or in traffic violations shall comply with Group 8 (commencing with Section 1215) of Subchapter 1 of Chapter 2 of Division 1 of Title 17 of the California Code of Regulations . . . (Health & Saf. Code, § 100700, italics added.) At the time of trial in this matter the applicable federal regulation (72 Fed.Reg. 71480, 71481-71483 (Dec. 17, 2007) [“Conforming Products List of Evidential Breath Alcohol Measurement Devices”]) listed both of the breath-testing devices used in this case—the Intoximeter Aleo-Sensor IV and the Intoximeter EC/IR—among 139 conforming evidential devices. The current version of the federal regulation does so as well. (77 Fed.Reg. 35746, 35748-35751 (June 14, 2012) [listing 140 conforming devices for evidential breath measurement].) Therefore, although the Intoximeter Aleo-Sensor IV was employed in this case as a preliminary screening device (see ante, fn. 1), it is in fact certified for evidential use. (See 2 Defense of Drunk Driving Cases, supra, ch. 19 [discussing and describing in detail both machines used in this case].) The statute reads in full: “All laboratories that are subject to the requirements of Section 100700 [(quoted ante, fit. 22)] shall ensure that breath alcohol instruments and calibrating devices used in testing are listed in the conforming products list in the Federal Register by the National Highway Traffic Safety Administration of the United States Department of Transportation.” (Health & Saf. Code, § 100701.) As noted earlier, at the time of trial in this case, and currently, the federal regulation lists both of the breath-testing machines used in this case. Defendant further argues that “the appearance of a breath machine on a federal approved list cannot prevent the defendant from challenging the accuracy of the machine involved in this case when he can show noncompliance with the regulations and unreliability.” In doing so the People reiterate their argument, set forth in their initial briefs, that there exist (1) a “broad legal definition” of “partition ratio” and (2) a “narrow scientific definition” of that concept—and they assert that this case can and should be resolved by analyzing the challenged testimony accordingly. The People’s argument does not appear to address the specific aspect of the evidentiary question that we consider here in part ILLA, of our analysis—namely, whether a defendant may present, with regard to the statutory per se offense, expert testimony that breath-testing machines are unreliable because inhaled air becomes saturated with alcohol from the upper airways and hence the machines fail to analyze any alveolar, deep lung air, despite the requirement of California Code of Regulations, title 17, section 1219.3 that “essentially” such air be sampled and analyzed. The federal regulation enumerates specific and highly technical “Test Methods and Requirements” for (1) “Precision and Accuracy”; (2) “Acetone Interference”; (3) “Blank Reading”; (4) “Breath Sampling” using a BASS device (see BASS for Qualification Testing of Breath Alcohol Measurement Devices, supra, described ante, at the end of pt. III.A.l.); (5) “Input Power”; (6) “Ambient Temperature”; (7) “Vibration Stability”; and (8) “Electrical Safety Inspection.” (58 Fed.Reg. 48705, 48707-48708.) Our conclusion is consistent with observations made by the appellate court in Borger v. Department of Motor Vehicles (2011) 192 Cal.App.4th 1118 [ 121 Cal.Rptr.3d 816 ], a civil administrative law decision not cited in the briefs. In that case, the plaintiff’s driver’s license was suspended under section 13557, subdivision (b)(2) (person under 21 driving with 0.08 percent or more, by weight, of alcohol in his or her blood). An expert witness testified for the plaintiff at the license-suspension hearing that the breath-testing machine used had a “plus or minus 0.02 percent” margin of error, and hence the plaintiff’s blood-alcohol concentration could have been “ ‘anywhere between a 0.06 and a 0.10.’ ” (Borger, supra, at p. 1121 .) The hearing officer did not credit this expert testimony, and suspended the license, but the trial court disagreed, and set aside the suspension. On review, the appellate court reversed. Although the Court of Appeal did not cite Health and Safety Code section 100701 or the relevant federal regulations, it did focus on the applicable state regulations, and noted, among other things, that “[a] challenge to the general reliability of an approved breath testing device is, in essence, a challenge to the regulation allowing the device to be on the approved list.” (Borger, supra, 192 Cal.App.4th at p. 1121 .) The court noted that the machine at issue was an “ ‘approved instrument’ within the meaning of California Code of Regulations, title 17, section 1221.3” and that under “[t]he logic of the trial court ruling” (ibid.) the expert’s testimony would “ ‘overrule’ every . . . reported result” of the breath-testing machine at issue “unless it is 0.10 percent or more. This would change the California Code of Regulations, title 17, sections 1221 through 1221.5 and effectively remove this breath testing device from the . . . ‘approved instrument’ list.” (Id., at p. 1122.) The court concluded that under Evidence Code section 664, results from approved and properly operated breath-testing machines are presumed to be in “ ‘compliance with statutory and regulatory standards, which in turn gives rise to an inference of reliability.’ ” (Borger, supra, at p. 1122 .) Accordingly, the appellate court ruled, the statutes and regulations required reversal of the trial court’s judgment. (Id., at p. 1123.) Likewise when the trial court asked Dr. Hlastala whether gender might also be a factor in his analysis, he agreed that it could be and stated, “There are some differences related to gender in the context of lung volume. Women have a smaller lung, therefore a higher breath test relative to that blood.” As alluded to earlier, the Arizona Court of Appeals appears to have come to a different conclusion in Cooperman, supra, 282 P.3d 446 . In that case the defendant was charged, as here, under both generic and per se statutes. The Arizona per se statute, like ours, defines the offense in terms of an impermissible blood or breath concentration. (Id., at p. 449.) The appellate court in Cooperman reaffirmed and extended prior rulings that testimony concerning partition ratio variability was properly admitted with regard to the generic count (id., at pp. 451-454), but properly excluded with regard to the breath test offered in support of the per se charge (id., at p. 454). Nevertheless, the court agreed with the defendant that very similar evidence—concerning “the possible effect on breath tests of hematocrit, breathing patterns, and breath and body temperature” (ibid.)—was separately “relevant to both charges” (that is, the generic and the per se counts)—because that evidence implicated “the ability of the machine to accurately register a defendant’s breath alcohol concentration.” (Id., at p. 455, italics added.) For the reasons set forth above, we decline to construe California law as the Cooperman court appears to have construed Arizona law. The apparently contrary conclusion reached by the Arizona Court of Appeals in Cooperman, supra, 282 P.3d at pages 454-455, appears distinguishable to the extent that court did not consider the matter from the perspective of deference to legislative determinations. In any event, as we observed ante, footnote 30, we find this aspect of Cooperman unpersuasive with respect to the evidentiary question that we address under California law. Finally, defendant asserts that exclusion of the proposed expert testimony “constituted a due process denial of a fair trial as well as denying the federal constitutional right to compulsory process.” This claim, which is evidently premised on asserted error under state law, fails because there was no state law error.