OPINION
{1} The opinion heretofore filed in this case is withdrawn and the following substituted therefor.
{2} This appeal examines the evidence required to sustain a conviction under the so-called per se section of the driving while under the influence (DWI) statute, NMSA 1978, § 66-8-102(0) (1999). We hold that a blood or breath alcohol test administered over two hours after the time of driving, and yielding only marginal results, must be corroborated by additional evidence to support a jury verdict. That corroborative evidence may take various forms which we do not attempt to catalogue in this opinion. The evidence might include a police officer’s observation of significant incriminating behav*707ior on the part of the driver, or the evidence might include expert testimony relating the test result back in time to the time of driving. However, in the absence of probative corroborating evidence, Defendant’s conviction lacks substantial evidence and cannot stand. Accordingly, we reverse Defendant’s conviction under Section 66-8-102(0). We affirm Defendant’s remaining conviction.
BACKGROUND
{3} At about 12:30 a.m., New Mexico State Police Officer T.E. Christian observed Defendant driving his vehicle across the median on 1-40 near Moriarity onto the interstate heading in the opposite direction. The officer stopped Defendant’s car because, in his opinion, it was illegal to turn in a median.
{4} Other than the illegal turn, the officer did not observe Defendant driving erratically or in an unsafe manner. Defendant stopped his vehicle promptly upon seeing the officer’s signal. At that point, the officer did not suspect Defendant of DWI. Upon conversing with Defendant, the officer detected a moderate odor of beer on his breath. His speech was normal. According to the officer’s testimony, Defendant admitted that during the past hour he had drunk a “few” beers or a “couple” of beers at a nearby bar in Moriarity. The officer then required Defendant to perform three field sobriety tests: the finger counting test, the counting backward test, and the horizontal gaze nystagmus (HGN) test which, according to the officer, Defendant failed. The officer then administered a breath alcohol test to Defendant. Defendant was charged with two DWI counts, under Sections 66-8-102(A) and 66-8-102(0), as well as driving illegally on a divided highway (crossing the median) in violation of NMSA 1978, Section 66-7-319 (1978). Approximately two hours and fifteen minutes after Defendant’s arrest, a blood alcohol test was administered which showed a blood alcohol concentration (BAC) of 0.08.
{5} At trial, the State first attempted to introduce the results of the breath alcohol test, but the State was unable to provide an adequate foundation for the accuracy of the test machine, and the court sustained Defendant’s objection. Those test results are not part of this record. Over Defendant’s objeetion, the State was then allowed to introduce the results of the blood alcohol test. At the conclusion of the State’s case, Defendant moved for a directed verdict based on the insufficiency of the evidence. The court granted Defendant’s motion for a directed verdict with respect to the DWI charge under Section 66-8-102(A) (“It is unlawful for any person who is under the influence of intoxicating liquor to drive____”). However, the court denied the motion with respect to the DWI charge under Section 66-8-102(0) (“It is unlawful for any person who has an alcohol concentration of eight one-hundredths or more in his blood or breath to drive____”). After Defendant presented his case, the jury deliberated and convicted Defendant of violating the per se section of the DWI statute, Section 66-8-102(C), as well as driving illegally on the median.
{6} Defendant raises three issues on appeal: (1) whether the trial court erred in denying Defendant’s motion for a directed verdict and post-trial motion directed to the DWI charge under Section 66-8-102(0) based upon the State’s failure to establish an evidentiary nexus between the blood alcohol test administered over two hours after driving and Defendant’s likely BAC at the time of driving, (2) whether the court erred in denying Defendant’s motion for a directed verdict on the charge of driving on divided highways under Section 66-7-319, and whether that conviction was supported by substantial evidence, and (3) whether the court erred in admitting into evidence the results of the HGN test. We affirm on the last two issues with minimal discussion. The first issue, however, presents a more substantial challenge, and is to that question that we now turn.
DISCUSSION
Defendant’s conviction under Section 66-8-102(0 was not supported by substantial evidence.
{7} A conviction under Section 66-8-102(C) requires proof of a BAC of 0.08 or more at the time of driving. New Mexico’s UJI 14-4503 NMRA 2001 sets out the essential elements for the State to prove: first, “[t]he defendant operated a motor vehicle,” and second, “[a]t that time, the defendant *708had an alcohol concentration of eight one-hundredths (.08) grams or more in [one hundred milliliters of blood] [or] [two hundred ten liters of breath].”
{8} Timing is an essential element of the crime. The State must prove a nexus between a BAC of 0.08 or more and the time “defendant operated a motor vehicle.” Id.; see also Biemer v. State ex rel. Taxation & Revenue Dep’t, 113 N.M. 696, 698, 831 P.2d 995, 997 (Ct.App.1992) (distinguishing a DWI charge from a driver’s license revocation proceeding in which the state does not have to prove a nexus between an excessive BAC and the time of driving). Timing is also an essential element for a conviction under Section 66-8-102(A). See UJI 14-4501 NMRA 2001. Because Defendant’s BAC test was administered over two hours after he finished driving, Defendant argues that the essential element of timing or relation back to the time of driving required direct proof, and that in the absence of such evidence the State left the element of timing to jury speculation.
{9} This is not our first occasion to address the “relation back” issue. In State v. Cavanaugh, 116 N.M. 826, 829-30, 867 P.2d 1208, 1211-12 (Ct.App.1993), we rejected a similar argument based on facts and circumstances specific to that case. There, we affirmed a DWI conviction under Section 66-8-102(C) when the police administered a blood alcohol test slightly less than two hours after the arrest which yielded a BAC of 0.13. We described that BAC as “sufficiently over the [legal] limit,” which was then set by statute at 0.10. Id. at 830, 867 P.2d at 1212.
{10} Significantly, the court in Cavanaugh noted that the BAC of 0.13 was “combined with the evidence of [defendant’s behavior before, at, and after the time of driving” to sustain the conviction. Id. After being detected with liquor on his breath, Cavanaugh fled the scene leading the police on a high speed chase. Id. at 829, 867 P.2d at 1211. Cavanaugh violently resisted arrest even afier his capture. Id. We held that those two kinds of evidence — a BAC sufficiently over the legal limit combined with egregious, incriminating behavior — met a threshold level of substantial evidence, and the jury was allowed to infer without direct evidence that Cavanaugh’s BAC was likely over the legal limit at the time of driving. Id. at 830, 867 P.2d at 1212.
{11} The specific facts on which we relied in Cavanaugh demonstrate just how different the case is before us. Here, in contrast to Cavanaugh, we have no BAC “sufficiently over the limit.” After the passage of over two hours, Defendant’s BAC was exactly at 0.08 and no more. Even the State’s expert acknowledged that a certain margin of error in the test results must be taken into account.1 Based on the observations of the arresting officer, Defendant’s contemporaneous behavior appears unremarkable when compared to Cavanaugh. Cf. id. (noting Cavanaugh’s egregious behavior “before, at, and after” the time of driving).
{12} Like Cavanaugh, Defendant had trouble with field sobriety tests. When requested to count backward from ten to one, Defendant continued to zero instead of stopping at one. When he was asked to count his fingers, Defendant omitted his pinky finger; then after re-instruction he counted correctly and then incorrectly. When given the HGN test, Defendant was unable to keep his head still despite instructions to do so. In regard to all three tests, the officer testified that Defendant’s inability to follow instructions meant that he had failed each test. The officer concluded that the results of these field sobriety tests, coupled with evidence of drinking, suggested that Defendant was under the influence of alcohol. The officer offered no testimony as to whether the field sobriety tests could be used to corroborate Defendant’s 0.08 BAC reading and relate it to the time of driving.
*709{13} We make one important observation in regard to how this evidence compares to Cavanaugh. Here, the field sobriety tests and other evidence of Defendant’s behavior might have been probative of the DWI charge under Section 66-8-102(A), but the trial court found that evidence so unpersuasive that it directed a verdict against the State on that charge. We think this is significant because, at least on this record, it neutralizes any inference one might otherwise try to draw from Defendant’s behavior to corroborate the BAC readings, relate them back to the time of driving, and prove a case under Section 66-8-102(C). In contrast, the court in Cavanaugh allowed the same charge under subsection (A) to go to the jury, and we upheld the sufficiency of the evidence. Id. at 829, 867 P.2d at 1211.
{14} We need not decide today whether, and under what circumstances, field sobriety tests results could be used to draw a nexus between a delayed BAC test result and the likely BAC at the time of driving. Today, we need only observe that no witness testified in this case as to any such nexus, and there was no other evidence on point. Therefore, we conclude under these facts that the field sobriety tests did not tend to make the alleged fact any more or less probable that Defendant had a BAC at or greater than 0.08 at the time he drove his car. See Rule 11-401 NMRA 2001.
{15} Without evidentiary support for a reasonable jury inference of the kind described in Cavanaugh, the jury was left to speculate about what the BAC might or might not have been at the time Defendant drove his vehicle. Without evidentiary guidance, no jury of lay persons can know, beyond a reasonable doubt, whether a 0.08 BAC test result would equate to a BAC reading at the time of driving of 0.08, 0.12, 0.04, or anything in between.
{16} Jurors share common human experience, and they are entitled to draw upon that experience to make reasonable inferences at trial. See State v. Mann, 2000-NMCA-088, ¶ 91, 129 N.M. 600, 11 P.3d 564 (stating that jurors are expected to rely upon them life experiences and background during deliberations); Duke City Lumber Co. v. N.M. Envtl. Imp. Bd., 95 N.M. 401, 405, 622 P.2d 709, 713 (Ct.App.1980) (stating that jurors may draw inferences from the trial evidence). Most jurors have some understanding of human behavior during a state of intoxication. For example, if an accused is observed falling down drunk, experience tells us that for most people such a condition likely arose over a period of time and not all of a sudden. Similarly, human experience guides us in deciding whether such an accused likely had the ability to drive an automobile in a prudent manner within a reasonable time before or after his behavior is observed in a state of intoxication. See Sanchez v. Wiley, 1997-NMCA-105, ¶ 19, 124 N.M. 47, 946 P.2d 650 (teenager competent to offer lay opinion about whether defendant driving under the influence of alcohol).
{17} But a BAC reading from a laboratory test is just a sterile number; by itself it tells us nothing about a driver’s condition hours earlier. Extrapolating backward in time can be a difficult task even for experts. See People v. Victory, 166 Misc.2d 549, 631 N.Y.S.2d 805, 811 (Crim.Ct.1995) (“ ‘[T]he longer the delay between the time of [the] incident and [the] sample collection, the more difficult it becomes, scientifically, to draw reasonable inferences from one ‘data point,’ back to the ‘driving’ time.’” (Quoting E. Fitzgerald & D. Hume, Intoxication Test Evidence: Criminal and Civil § 2:30 [Lawyers Co-op 1987 & 1994 Supp.].)). For a lay jury, unassisted by corroborating evidence, the assignment becomes mere camouflage for guesswork.
{18} Basing a criminal verdict upon such conjecture ignores the standard of proof— beyond a reasonable doubt — and undermines the most basic principles of fairness. See State v. Garcia, 114 N.M. 269, 275, 837 P.2d 862, 868 (1992) (“[E]vidence equally consistent with two inferences does not, without more, provide a basis for adopting either one — especially beyond a reasonable doubt.”); see also State v. Gallow, 185 Ariz. 219, 914 P.2d 1311, 1313 (Ct.App.1995) (holding that when the state’s criminalist could not testify that the defendant’s BAC at the time of driving exceeded statutory limit beyond a reasonable doubt, the jury was not *710permitted to make the inference that it did); Haas v. State, 597 So.2d 770, 775-76 (Fla. 1992) (Kogan, J., concurring in part, dissenting in part) (To allow the state to presume the defendant’s BAC at the time he was driving from test results several hours later, would relieve the state of its burden to prove guilt beyond a reasonable doubt, especially as scientific and medical evidence contradicts the notion. The presumption was “contrary to the most basic conception of due process embodied in ... the Fourteenth Amendment.”); Commonwealth v. Barud, 545 Pa. 297, 681 A.2d 162, 166 (1996) (holding that a statute that permitted a conviction with a BAC taken three hours after a person drove was unconstitutional because it “completely fails to require any proof that the accused’s blood alcohol level actually exceeded the legal limit at the time of driving ”).
{19} We offer one critical observation. As a matter of sound public policy, our legislature could choose to create a statutory inference that a 0.08 BAC within a specified time, say two or three hours after driving, is prima facie evidence of a per se violation of Section 66-8-102(C), which a defendant could then try to rebut. See Bierner, 113 N.M. at 698, 831 P.2d at 997. Other state legislatures have chosen this course of action. See Haas, 597 So.2d at 774 (discussing legislative schemes in Minnesota and California); Barud, 681 A.2d at 167 n. 6 (collecting states with such criminal codes); cf. Allman v. State, 728 N.E.2d 230, 234 (Ind.Ct.App.2000) (holding that when the statutory time had been exceeded the state was required to relate the BAC back to the time of the driving by competent evidence); Victory, 631 N.Y.S.2d at 815 (holding that “the People must now prove at a hearing by expert testimony the scientific reliability of such BAC test administered more than two hours from arrest”). We emphasize that although our legislature has the authority to make such a reasoned policy judgment, a lay jury does not. The duty of a jury is to apply the law, not to make it.
{20} State courts vary considerably in them approach to how per se violations of drinking and driving laws are proven. Only a few demand scientific relation-back evidence in all situations. Some courts uphold convictions based solely on subsequent BAC readings because, as previously mentioned, them statutes specifically authorize a relation-back inference if the BAC reading is taken within two or three hours after driving. See, e.g., State v. Guerra, 191 Ariz. 511, 958 P.2d 452, 456 (Ct.App.1998) (holding that “relation-back evidence was not required in order for the jury to be instructed on the presumption that defendant was under the influence of intoxicating liquor”); State v. Wetzel, 7 Haw.App. 532, 782 P.2d 891, 895 (1989) (same); State v. Ulrich, 17 Ohio App.3d 182, 478 N.E.2d 812, 822 (1984) (same). In those states without a statutory relation-back period, some court opinions require scientific testimony to relate the BAC back to the time of driving, while others do not. Compare Desmond v. Superior Court, 161 Ariz. 522, 779 P.2d 1261, 1267 (1989) (in banc) (“relating the BAC back to the time of arrest will most often require testimony by someone qualified as an expert”), supersession by statute recognized in Guerra, 958 P.2d at 455; State v. Rollins, 141 Vt. 105, 444 A.2d 884, 886-87 (1982) (requiring relation-back evidence) with Haas, 597 So.2d at 774 (not requiring relation-back evidence).
{21} In other jurisdictions, the BAC creates a rebuttable presumption of DWI, with the relation-back question left to the weight, not the admissibility of the evidence, as long as the test is taken within a reasonable period of time. See State v. Kubik, 235 Neb. 612, 456 N.W.2d 487, 501 (1990); State v. Taylor, 132 N.H. 314, 566 A.2d 172, 175-76 (1989). At least one state makes a-BAC, obtained within a reasonable time after the defendant was driving, a conclusive presumption of DWI and expressly prohibits the introduction of relation-back evidence by either the prosecution or defense. See State v. Tischio, 107 N.J. 504, 527 A.2d 388, 397 (1987).
{22} Other states, like New Mexico, allow convictions based on a subsequent BAC, and without scientific relation-back evidence, if the BAC is sufficiently corroborated by additional evidence of aberrant behavior on the part of the accused. See People v. Kappas, 120 Ill.App.3d 123, 76 Ill.Dec. 1, 458 N.E.2d 140, 144 (1983); State v. Ladwig 434 N.W.2d *711594, 595 (S.D.1989) (“ ‘[I]t is not necessary for an expert witness to extrapolate for the jury. However, ... evidence necessary to enable the court and/or the jury to extrapolate is required____(Quoting State v. McDonald, 421 N.W.2d 492, 494 n. 2 (S.D. 1988).)).
{23} New Mexico jurisprudence has already started down the path chosen in Cavanaugh, and we intend to stay the course. New Mexico has indicated that with sufficient, corroborative evidence a jury may reasonably infer that an excessive BAC reading relates back to the time of driving. That evidence is lacking in this ease.
{24} Accordingly, we conclude that the State’s evidence against Defendant was insufficient as a matter of law to sustain a guilty verdict for the per se DWI offense under Section 66-8-102(0). Having reversed Defendant’s conviction, we need not address his appellate argument that he was entitled to a directed verdict at the close of the State’s ease. See State v. Vargas, 42 N.M. 1, 4, 74 P.2d 62, 64 (1937) (“This rule is as firmly established in this jurisdiction as the Rock of Gibraltar.” Once a defendant chooses to present evidence, he waives appeal of the directed verdict.).
Defendant’s conviction for driving on divided highways, Section 66-7-319, is affirmed.
{25} Section 66-7-319, driving on divided highways, reads:
Whenever any highway has been divided into two roadways by leaving an intervening space or by a physical barrier or clearly indicated dividing section so constructed as to impede vehicular traffic, every vehicle shall be driven only upon the right-hand roadway and no vehicle shall be driven over, across or within any such dividing space, barrier or section, except through an opening in such physical barrier or dividing section or space or at a crossover or intersection established by public authority.
{26} The State’s evidence established that Officer Christian observed Defendant driving his vehicle across the median separating the easterly and westerly roadways on Interstate 40. The officer further testified that this occurred at a site that was not a duly authorized crossover or intersection. Defendant allegedly confessed to the officer that he “knew he wasn’t supposed to cross the median.”
{27} There was a pathway of some kind at or near where Defendant crossed. The testimony conflicted regarding whether Defendant actually used the pathway or just drove over the grass near the area of the pathway. Defendant argues on appeal that under the literal wording of the statute this pathway constituted an “opening” in the dividing space through which Defendant could legally cross, whether or not it was also a “crossover ... established by public authority.” Section 66-7-319. Defendant bases his argument exclusively on the wording of the statute.
{28} We need not address Defendant’s reading of the statute. The jury was instructed, without objection, on the following elements of the offense:
1. The defendant was driving his vehicle on a highway divided into two roadways.
2. The defendant drove his vehicle over, across or within any dividing space, on section dividing the two roadways.
3. When so doing, the defendant did not use a crossover established by public authority.
4. This happened in Torrance County, State of New Mexico, on or about the 12th day of November, 1997.
{29} The jury instruction becomes the law of the case, and Defendant does not challenge this instruction on appeal. See State v. Riddall, 112 N.M. 78, 84, 811 P.2d 576, 582 (Ct.App.1991). Under this instruction, the State had to prove only that Defendant drove across “any dividing space” and “did not use a crossover established by public authority.” Apparently, Defendant never argued for an instruction that would include his interpretation of the statute, and we are not referred to any such request below. Based on the officer’s testimony, there was substantial evidence before the jury to support a conviction for the offense as defined in the jury instruction.
*712{30} Defendant also argues on appeal that he was entitled to a motion for a directed verdict on the basis of this same challenge to the evidence in light of the language in the statute. After his motion was denied, Defendant proceeded to present evidence in his defense. It is well-settled that a defendant who presents evidence “waive[s] his claim that the evidence at the close of the State’s case was insufficient for submission to the jury.” State v. Lard, 86 N.M. 71, 73, 519 P.2d 307, 309 (Ct.App.1974). Defendant’s directed verdict argument is frivolous. See Vargas, 42 N.M. at 4, 74 P.2d at 64.
Admission of the HGN test results.
{31} Citing State v. Torres, 1999-NMSC-010, 127 N.M. 20, 976 P.2d 20, Defendant challenges the admission of evidence pertaining to the HGN test that Officer Christian administered to him at the scene. Defendant questions the foundation for the scientific evidence offered, as well as the officer’s lack of experience, training, and education to testify regarding the reliability of the test. See State v. Alberico, 116 N.M. 156, 861 P.2d 192 (1993).
{32} There is no need to discuss Defendant’s objections in this opinion. Insofar as the HGN test results were admitted to prove DWI under Section 66-8-102(A), Defendant’s objections were rendered moot when the trial court directed a verdict on that charge. Because we are reversing Defendant’s conviction for DWI under Section 66-8-102(0) for other reasons, we need not address the propriety of the trial court admitting the HGN test into evidence.
CONCLUSION
{33} We affirm Defendant’s conviction for violating Section 66-7_IS19, driving on divided highways. We reverse Defendant’s conviction for DWI in violation of Section 66-8-102(C).
{34} IT IS SO ORDERED.
WE CONCUR: M. CHRISTINA ARMIJO and CYNTHIA A. FRY, Judges.