dissenting in part.
AS 28.35.0380 is the statute that defines the offense of driving under the influence and that specifies certain procedures and penalties pertaining to that offense. In 2004, in response to this Court's decision in Conrad v. State, the Alaska Legislature made two major changes to the DUI statute. For the reasons explained in this dissent, I conclude that one of the legislature's changes is constitutional, but the other is not.
Background of this case: this Court's construction of the DUI statute in Conrad
AS 28.35.080(a) defines two ways in which the operator of a motor vehicle can commit the offense of driving under the influence. Under subsection (a)(1), a motorist commits this crime by being "under the influence" of alcohol or controlled substances. Under subsection (a)(2), a motorist commits this crime by having a blood aleohol level of .08 percent or higher.
Subsection (a)(1), the "under the influence" clause, has always been interpreted to require proof that a motorist was under the influence of intoxicants at the time of driving (as opposed to being under the influence either before or after driving). However, there was a debate as to whether, under subsection (a)(2), a motorist could be convicted of DUI if their blood aleohol level was .08 percent or greater at the time they submitted to the chemical test, even though their blood alcohol level might have been lower at the time of driving.
We resolved this debate in Conrad v. State, 54 P.3d 313, on rehearing 60 P.3d 701 (Alaska App.2002). In Conrad, we held that subsection (a)(2) required proof that a motorist's blood alcohol level equaled or exceeded .08 percent at the time of driving. If a later chemical test of the motorist's breath or blood showed that the motorist had a blood alcohol level of .08 percent or more, this created a presumption that the motorist's blood aleohol level was at least that high at the time of driving-but this presumption was rebuttable, and the motorist was entitled to introduce evidence tending to show that the result of the chemical test was not an accurate indicator of the motorist's blood alcohol level at the time of driving. Conrad, *34954 P.3d at 315-16, on rehearing 60 P.3d at 702.
The legislature's response to Conrad: the 2004 amendments to the DUI statute
In 2004, the legislature responded to the Conrad decision by making two significant changes to the DUI statute.
First, the legislature rewrote subsection (a)(2), the subsection that defines the "blood alcohol level" theory of the crime. Conrad held that a person violated subsection (a)(2) if they operated a motor vehicle at a time when their blood alcohol level was .08 percent or higher. In response to Conrad, the legislature amended subsection (a)(2) so that, now, a person violates subsection (a)(@2) if they operate a motor vehicle and if, within four hours of their operation of the vehicle, their blood aleohol level is .08 percent or higher, and if this blood alcohol level is attributable to the person's voluntary consumption of alcoholic beverages either before or during their operation of the vehicle.
(I agree with my colleagues that this amended version of subsection (a)(2) is constitutional, and I reach that conclusion for the reasons set forth in the majority opinion.)
Second, the legislature enacted a new subsection, AS 28.35.080(s), that prohibits defendants from raising a "delayed absorption" defense in a DUI prosecution.1
The "delayed absorption" defense is premised on the fact that there is an inevitable delay between a person's consumption of alcoholic beverages and their body's absorption of the aleohol into their bloodstream. This means that a person's blood aleohol level can continue to rise for some amount of time after they stop drinking. Because of this delay in the body's absorption of alcohol, there will be occasions when a chemical test administered to an arrested driver will yield a result that is significantly higher than the person's blood alcohol level at the time they were driving.
This phenomenon is illustrated by the facts of a case recently decided by this Court: Painter v. State, Alaska App. Memorandum Opinion No. 5192, 2007 WL 706752 (March 9, 2007).
When the defendant in Painter was initially stopped by the state troopers, he submitted to a preliminary breath test which showed his blood aleohol level to be .072 percent. (That is, the test showed that his blood alcohol level, according to the PBT, was under the legal limit.) One hour later, at the trooper station, Painter submitted to a second breath test, this time on a DataMas-ter; this second test showed his blood aleohol level to be .082 percent (i.e., slightly over the legal limit). Thirty-five minutes later, Painter obtained an independent blood test, and the result of this third test was .092 percent (i.e., considerably over the legal limit).
This potential discrepancy between a person's blood alcohol level at the time of testing and their blood alcohol level at the time they were driving is now irrelevant in prosecutions under subsection (a)(2) of the statute-because subsection (a)(2) now declares that a person's guilt hinges on the test result, and not on the person's blood alcohol level at the time of driving.
But this potential discrepancy remains important in prosecutions under subsection (a)(1) of the statute. The legislature did not amend subsection (a)(1); this subsection still requires proof that a person operated a motor vehicle "while under the influence of an alcoholic beverage". Thus, to prove the offense of DUI under subsection (a)(1), the government must establish that the defendant operated or otherwise controlled a motor vehicle at a time when the defendant's ability to control the vehicle was impaired *350due to the consumption of alcohol.2 It is not sufficient to prove that the defendant operated a motor vehicle and then, sometime later, became impaired.
Because, under subsection (a)(1), the ultimate fact to be proved is the defendant's impairment, a defendant's guilt or innocence does not directly hinge on the result of the defendant's post-arrest chemical test. Theoretically, a person might have a blood aleohol level of .08 percent (or greater) and still not be impaired. Conversely, as we recognized in Ballard v. State, a person might be "under the influence" for purposes of subsection (a)(1)-that is, the person might be demonstrably impaired-even though their blood alcohol level is less than the legal limits3
Nevertheless, the result of the defendant's post-arrest chemical test can be an important element of the State's proof in prosecutions under subsection (a)(1). This is because of the evidentiary presumptions codified in a companion statute, AS 28.35.-088(a).
In AS 28.35.033(a)(8), the legislature has decreed that when a person's post-arrest chemical test result is .08 percent blood alcohol or greater, this creates a rebuttable presumption that the person was "under the influence of an aleoholic beverage" for purposes of prosecutions under subsection (a)(1). In other words, in prosecutions under subsection (a)(1), if the defendant's post-arrest chemical test yielded a result of .08 percent or higher, a jury may lawfully find (based on this evidence alone) that the defendant was "under the influence" at the time that the defendant operated the motor vehicle.
(See Doyle v. State, 633 P.2d 306, 310-11 (Alaska App.1981), where we construed this statute to create a backward-looking presumption: a presumption that authorizes the jury to find, based on the result of a defendant's post-arrest chemical test, that the defendant was under the influence at the time of driving.)
The statute challenged in this appeal-AS 28.935.030(s)-was enacted for the purpose of limiting the ways in which a defendant can rebut this statutory presumption of impairment. Specifically, this new subsection declares that a defendant is barred from introducing evidence of delayed absorption to rebut the presumption of impairment that arises under AS 28.35.033(a2)(8) if the result of the defendant's post-arrest chemical test is .08 percent or higher.
Subsection (s) reads:
(s) In a prosecution [for driving under the influence], ... [evidence of] the [defendant's] consumption of aleohol before operating or driving may not be used as a defense that the [defendant's] chemical test did not [reflect] the [defendant's] blood aleohol at the time of the operating or driving. [However, evidence of the defendant's] [clonsumption of aleohol after operating or driving the motor vehicle, aircraft, or watercraft may be used to raise such a defense. [Emphasis added]
In other words, subsection (s) prohibits a defendant from introducing evidence that, because of the natural delay in the absorption of aleohol into the bloodstream, the defendant's blood alcohol level may have risen significantly between the time that the defendant stopped driving and the time that the defendant took the post-arrest chemical test-and that, consequently, the defendant's post-arrest chemical test result might not be a trustworthy indicator of the defendant's level of intoxication at the time that the defendant was operating the motor vehicle.
Why I conclude that subsection (s) is unconstitutional when applied to prosecutions for DUI under subsection (a)(1)
In prosecutions under subsection (a)(1), the ultimate question is whether the defendant was "under the influence" at the time that they operated the motor vehicle A defendant's post-arrest chemical test result does not directly prove or disprove this ultimate fact. However, because of the eviden-tiary presumption codified in AS 28.35.033(a)(8), the government can rest its case solely on the defendant's chemical test result. Pursuant to this statute, juries are *351instructed that if a defendant's post-arrest chemical test result was .08 percent or higher, the jury can find that the defendant was impaired at the time of driving.
It is undisputed that there is a delay between a person's consumption of alcoholic beverages and the absorption of the alcohol into the person's bloodstream (where it causes intoxication). This delay in the absorption of alcohol into the bloodstream is obviously relevant to the issue of whether a defendant's post-arrest chemical test result is an accurate indicator of whether the defendant was impaired earlier, at the time of driving. Nevertheless, subsection (s) prevents defendants from offering evidence of delayed absorption to contest the statutory presumption of impairment that arises when a defendant's post-arrest chemical test result is .08 percent or higher.
This preclusion of evidence is irrelevant in prosecutions under subsection (a)(2)-because, under the amended version of subsection (a)(2), the ultimate fact to be proved is now the defendant's blood aleohol level at the time of the chemical test, and not the defendant's earlier blood alcohol level at the time of driving. It therefore does not matter if the defendant's blood alcohol level might have been lower at the time of driving.
But in prosecutions under subsection (a)(1), the government must prove that the defendant was impaired at the time of driving. It is not enough to show that the defendant operated a motor vehicle and then became impaired later. Because of this, the rule of evidence preclusion codified in subsection (s) violates the guarantee of due process of law when it is applied to prosecutions under subsection (a)(1). It unjustifiably prevents defendants from introducing evidence that is both scientifically valid and directly relevant to the question of whether the defendant was impaired by alcohol at the time of driving.
As our supreme court stated in Smithart v. State, "a defendant's right to present a defense is a fundamental element of due process." 988 P.2d 583, 586 (Alaska 1999). See also Keith v. State, 612 P.2d 977, 982-83 (Alaska 1980).
Both Smithart and Keith involved eviden-tiary rulings by a trial judge that improperly restricted the defendant's presentation of a defense. However, the same principle applies when the evidentiary restriction is enacted by the legislature,4 or is codified in a rule of court,5 or is adopted as a rule of common law by a state's appellate courts.6
I acknowledge that the legislature has broad discretion when defining the elements of crimes, and that this broad discretion includes the authority to define elements in a manner that renders certain matters irrelevant (and thus inadmissible). For instance, in Abruska v. State, 705 P.2d 1261, 1265-66 (Alaska App.1985), and Neitzel v. State, 655 P.2d 325, 334-35 (Alaska App.1982), this Court upheld the legislature's authority to define the culpable mental states of "knowingly" and "recklessly" in such a way as to preclude a potential defense of voluntary intoxication.
See also Montana v. Egelhoff, 518 U.S. 37, 56, 116 S.Ct. 2013, 2023-24, 135 L.Ed.2d 361 (1996) (upholding a state legislature's authority to preclude evidence of a defendant's voluntary intoxication to rebut the government's proof of the defendant's culpable mental state); Fisher v. United States, 328 U.S. 463, 475-77, 66 S.Ct. 1318, 1324-25, 90 L.Ed. 1382 (1946) (upholding Congress's authority to preclude a defendant from introducing evidence of a mental disability or deficiency short of insanity on the issue of the defendant's capacity for premeditation and deliberation); Muench v. Israel, 715 F.2d 1124, 1144-45 (7th Cir.1983) (holding that the states "[are] not constitutionally compelled to recognize the doctrine of diminished capacity", and that therefore "a state may exclude expert testimony offered for the purpose of establishing that a criminal defendant lacked the capacity to form a specific intent."); Peo*352ple v. Carpenter, 464 Mich. 223, 627 N.W.2d 276, 284-85 (2001) (same).
The Alaska Legislature pursued this course when it amended the "blood aleohol level" clause of the DUI statute, subsection (a)(2): the legislature amended the definition of the crime in a way that made it irrelevant whether the result of the defendant's post-arrest chemical test accurately reflected the defendant's blood alcohol level at the time of driving.
But the definition of DUI under subsection (a)(1) remains the same as it has always been: the government must prove that the defendant was under the influence at the time the defendant operated the motor vehicle.
Moreover, as we explained in Doyle, the evidentiary presumption codified in AS 28.35.083(a)(8) is a backward-looking presumption that relates to the defendant's physical and mental condition at the time of driving. Under this statute, a test result of .08 percent or higher does not give rise to a presumption that the defendant was under the influence at the time of the test; rather, it gives rise to a presumption that the defendant was under the influence at the time of driving.
Because the ultimate fact to be proved is the defendant's impairment at the time of driving, and because the scientifically recognized phenomenon of delayed absorption is directly relevant to the jury's assessment of whether the result of the defendant's post-arrest chemical test is a trustworthy indicator of the defendant's earlier impairment, the legislature may not prohibit the defendant from presenting evidence about delayed absorption, and may not prohibit the defendant from asking the jury to consider this phenomenon when the jury assesses whether the defendant was under the influence at the time of driving.
For these reasons, I conclude that the evidentiary prohibition codified in AS 28.35.080(s) violates a defendant's right to due process of law when the defendant is prosecuted for DUI under AS 28.35.080(a)(1) and the government relies on the result of a post-driving chemical test.
. In the 2004 legislative hearings, this defense was referred to as the "big gulp" defense-a reference to the possibility that a motorist had hurriedly consumed a large quantity of an alcoholic beverage shortly before the police pulled the motorist over. Although the phrase "big gulp" may be catchier than "delayed absorption", this latter phrase is a more accurate description of the defense. There are several potential reasons why the absorption of alcohol into a person's bloodstream might be delayed-for example, illness or the recent consumption of a heavy meal. And it is this delay that is the crucial component of any defense claim that a post-arrest chemical test result might not accurately indicate the defendant's blood alcohol level at the time of driving.
. Swensen v. Anchorage, 616 P.2d 874, 880 n. 10 (Alaska 1980); Ballard v. State, 955 P.2d 931, 940 (Alaska App.1998).
. Ballard, 955 P.2d at 940.
. See Davis v. Alaska, 415 U.S. 308, 311, 320, 94 S.Ct. 1105, 1108, 1112, 39 L.Ed.2d 347 (1974).
. See id.
. See Chambers v. Mississippi, 410 U.S. 284, 294-98, 93 S.Ct 1038, 1045-47, 35 L.Ed.2d 297 (1973).