Price, J., delivered the opinion of the Court, in which McCormick, P.J., and Meyers, Keller, Holland, Womack, Johnson, and Keasler, J.J., join. Mansfield, J., concurs in the judgment only on point three, and otherwise joins the opinion.

Appellant was convicted of capital murder on March 12, 1998. Tex. Penal Code Ann. §19.03(a). Pursuant to the jury’s answers to the special issues set forth in Texas Code of Criminal Procedure article 37.0711 §§3(b) and (e), the trial judge sentenced appellant to death. Art. 37.071 §3(g).¹ Direct appeal to this Court is automatic. Art. 37.0711 §3(j). Appellant raises eight points of error. We will affirm.

In his first two points of error, appellant challenges the legal sufficiency of the evidence to support his capital murder conviction for killing more than one person in the same criminal transaction. We review the evidence in the light most favorable to the verdict. Jackson v. Virginia, 443 U.S. 307, 319 (1979).

The evidence at trial revealed that the victims, Forrest Henderson and Richard Wrotenbury, were singers in the Houston Grand Opera. Shortly before his death, Henderson toured with the opera in Scotland. Wrotenbury moved into Henderson’s Houston apartment to housesit while Henderson was out of the country and continued to live in the apartment after Henderson returned.

David Trujillo and Roger Lindgroff² lived next door to Henderson and Wrotenbury. At around 10:30 p.m. on September 10, 1988, Trujillo heard music and Henderson’s voice through the common wall separating their apartments. Trujillo went to sleep around 2:00 a.m. and was awakened at 4:45 a.m. by the sound of Wrotenbury screaming several times, “Oh my God. No. No.” Trujillo also heard what sounded like someone being hit numerous times with a pipe or a baseball bat. After 30 minutes of silence, he heard the water running for about 45 minutes. Lindgroff started to knock on their neighbor’s door to see if there was a problem, but Trujillo called him back inside. Trujillo never heard Henderson’s front door open or anyone leave. A person could enter or leave Henderson’s apartment via a separate stairwell, however, without having to pass by Trujillo’s door.

Trujillo explained that, before Wrotenbury moved in, he would see “street trash” going in and out of Henderson’s apartment, that the apartment was a rowdy place, and that there was always some kind of screaming and fighting going on over there. Since Wrotenbury had moved in, however, the rowdiness had subsided.

Besides the opera, Wrotenbury also worked as a music teacher at Deer Park Elementary School; but on Monday, September 12, 1988, he failed to appear for work At 9:00 a.m., the school principal contacted Henderson’s apartment manager to check on him. The manager unlocked Henderson’s apartment door and found nothing disturbed in the living room and kitchen. He proceeded to one of the bedrooms, pushed open the door, and saw a body covered with blood. He promptly left and called 911.

Police officers arrived at the apartment soon thereafter and detected no signs of forced entry. They found Henderson’s and Wrotenbury’s bodies in their respective bedrooms at opposite ends of the apartment. Henderson’s nude body was lying face- down in his bed, and Wrotenbury’s body, clad only a pair of swimming trunks, was lying on the floor of his bedroom. Absence of significant blood in the hallway connecting the two bedrooms indicated that neither victim left his room during or after the attacks. Police found a bloody metal bar in the hallway and a bloody knife in the kitchen sink. Blood was all over the bedroom walls, doors, and curtains. Both victims’ wallets were missing, and Henderson’s car was gone.³

The forensic pathologist testified that Alan Wrotenbury suffered a severed carotid artery, cuts to the vertebrae, and at least three blows to the back of the head with a narrow blunt instrument consistent with a pipe. The force of one of the blows Wrotenbury received knocked out a tooth. Forrest Henderson had received a shallow, non-fatal cut to the neck, defensive wounds on both arms, a six-inch fracture of the skull from a blunt force, and multiple stab wounds to the torso. Fixed lividity in both bodies signified that both people had been dead for more than eight hours. Tests performed on both victims revealed no signs of drugs, alcohol, or semen.

Blood samples and 20 identifiable fingerprints were collected from the crime scene, but the Houston Police Department (“HPD”) was unable to develop a suspect. In 1995, HPD upgraded to a new fingerprint system with an expanded database. The new system matched appellant with prints lifted from a beer can and a glass tumbler in Henderson’s bedroom. A bloody print found on Henderson’s bedroom door also matched appellant. An expert in blood-spatter interpretation testified that the bloody fingerprint could have been formed only by touching a blood drop while the blood was still wet—as opposed to a blood drop landing on an old fingerprint.

An HPD serologist testified that type-B blood was found on a bedroom door. Appellant is blood-type B; both victims were blood-type A. Only these blood types were detected at the crime scene. The State’s DNA expert testified that appellant’s DNA profile matched DNA isolated from blood stains on a red towel and a beige towel located in Henderson’s bathroom. The odds that another African-American would possess the same profile is one in 7.2 million. Further, DNA analysis could not exclude appellant as a contributor of the blood mixture covering the metal bar.

In reviewing the sufficiency of the evidence, we must view the evidence in the light most favorable to the verdict and determine whether any rational trier of fact could have found the essential elements of the offense beyond a reasonable doubt. Jackson, 443 U.S. at 319. Although our evaluation considers all evidence presented at trial, Johnson v. State, 967 S.W.2d 410, 412 (Tex. Crim. App. 1998), we may not re-weigh the evidence and substitute our judgment for that of the jury. Wilson v. State, 863 S.W.2d 59, 65 (Tex. Crim. App. 1993).

In his first point of error, appellant alleges that the fingerprint and DNA evidence does not establish him as the perpetrator because he could have left prints on the beer can and tumbler before the killings. Appellant further contends that he could have reentered the apartment after the killings to leave the bloody fingerprint on the door and his own blood on the bathroom towels. Appellant’s scenario does not establish that the evidence presented at trial was insufficient to support the jury’s verdict.

Appellant’s bloody fingerprint puts him in the apartment while the blood was still wet. Further, the discovery of blood consistent with appellant’s DNA profile on the towels and on the bloody bar leads to the reasonable conclusion that appellant was injured during the struggle with one or both of the victims. Thus, the evidence is legally sufficient to find that appellant killed Wrotenbury and Henderson. Point of error one is overruled.

In his second point of error, appellant argues that the evidence is insufficient to establish that Henderson and Wrotenbury were killed during the course of the same criminal transaction. As support for this allegation, appellant relies on the testimony of (1) Trujillo that he did not hear Henderson’s voice after 10:30 p.m. and heard only Wrotenbury struggling several hours later and (2) the medical examiner that he could not pinpoint the time of death of either victim. Appellant argues that someone could have entered the apartment through the private staircase, killed Henderson, left the apartment, reentered the apartment, and killed Wrotenbury.⁴

To convict appellant of capital murder, the jury in this case was required to find that appellant intentionally caused the death of two individuals, Forrest Henderson and Alan Wrotenbury, “during the same criminal transaction.” Tex. Penal Code § 19.03(a)(6)(A) (Vernon Supp. 1988). Because the legislature did not define the term “same criminal transaction,” we have interpreted that phrase to mean “a continuous and uninterrupted chain of conduct occurring over a very short period of time . . . in a rapid sequence of unbroken events.” Rios v. State, 846 S.W.2d 310, 311-312 (Tex. Crim. App. 1992), cert. denied, 507 U.S. 1051(1993); Vuong v. State, 830 S.W.2d 929, 941 (Tex. Crim. App. 1992), cert. denied, 506 U.S. 997 (1992).

Here, both victims were killed in the same manner with the same weapons and were found dead in the same apartment. Although the medical examiner testified that only Henderson’s body showed signs of decomposition, he maintained that he could not determine who was killed first. He explained that Henderson’s struggle with the assailant could have caused increased secretion of acid in his system, thus accelerating decomposition. Based on our review of the record, we find that “the jury could rationally conclude appellant engaged in a continuous and uninterrupted process, over a short period of time, of carrying on or carrying out murder of more than one person.” Rios, 846 S.W.2d at 314. Appellant’s second point of error is overruled.

In his third point of error, appellant complains that the trial court denied him a hearing under Kelly v. State, 824 S.W.2d 568 (Tex. Crim. App. 1992) and Daubert v. Merrill-Dow Pharmaceuticals, 509 U.S. 579 (1993), after he challenged the reliability of the State’s DNA evidence. Specifically, appellant requested the court to determine whether “the DNA was collected and tested properly pursuant to accepted protocols, subjected to peer reviews, and properly calculated.” Appellant further points out that at trial he “also expressed concern over the match windows used in connection with RFLP⁵ DNA testing conducted by the state.”

We have stated numerous times that a trial court’s responsibility under Texas Rule of Criminal Evidence 702 is to determine whether proffered scientific evidence is sufficiently reliable and relevant to assist the jury.⁶ See, e.g., Kelly, 824 S.W.2d at 573; Jordan v. State, 928 S.W.2d 550, 554-55 (Tex. Crim. App. 1996). The proponent of the scientific evidence bears the burden of demonstrating by clear and convincing evidence that the evidence is reliable. This is accomplished by showing: (1) the validity of the underlying scientific theory; (2) the validity of the technique applying the theory; and (3) proper application of the technique on the occasion in question. Kelly, 824 S.W.2d at 573. We have emphasized that, before scientific evidence may be admitted, the trial court must conduct a hearing outside the presence of the jury to determine whether the proponent has established all three criteria. Id; Massey v. State, 933 S.W.2d 141, 152 (Tex. Crim. App. 1996); Campbell v. State, 910 S.W.2d 475, 478-79 (Tex. Crim. App. 1995), cert. denied, 517 U.S. 1140 (1996). This pre-admission determination is required whether the science at issue is novel or well established. Hartman v. State, 946 S.W.2d 60, 63 (Tex. Crim. App. 1997); cf. Kumho Tire Co., Ltd. v. Carmichael, 119 S.Ct. 1167, 1174 (1999) (holding that trial court’s “gatekeeping” function under Federal Rule of Evidence 702 applies to all expert testimony—not just scientific testimony).

In this case, defense counsel agreed to the general validity of DNA evidence under the first prong of Kelly but challenged the validity and application of the technique, under the second and third prongs of Kelly, used to identify appellant as a suspect. Before the State presented its DNA evidence, the court heard arguments from the State and appellant debating the reliability of the State’s evidence. Emphasizing that DNA evidence is based on a conceptually valid theory, the court informed the parties that it was not required under Daubert and Kelly to conduct a pre-admission hearing to determine whether HPD’s techniques were valid and properly performed in this case. The court interpreted the second two Kelly prongs as matters for the jury to decide:

[N]or am I persuaded that it is necessary for me to have a hearing to determine the specific manner and means in which collections occur, evaluations are conducted.

Certainly, that is permissible for you to present to a jury for the purposes of refuting whatever it is that the State claims. But it is not the matter of a subjective pre-trial hearing for the purposes of suppressing it because if there is a dispute, the resolution of the dispute must be done by the jury based on the credibility of the witnesses.

The defense then requested to call its expert to discuss the State’s DNA evidence:

I’m going to call Dr. Elizabeth Johnson and I’m going to ask her, specifically, based on the evidence given to us under the discovery orders, are there some specific protocols that are either miscalculated regarding the match windows and are those subject to peer review.

And, specifically, I’m going to ask her under the RFLP protocol, some specific questions regarding our discovery and our evidence that was given to us of the specific problems.

And then under the PCR, are there some specific problems. And that will be it.

Before defense counsel could pursue these issues, however, the court intervened to question the defense’s expert when she took the stand:

THE COURT: Dr. Johnson, have you seen or had a chance to read whatever information [defense counsel] has supplied to you regarding this stuff?

THE WITNESS: Yes, I have.

THE COURT: From what you have seen, can you specifically say that there was anything done in terms of the collection, the keeping, the analysis, the conclusions, can you say from what you have, specifically, that there was any conduct that was incompatible with what you would say to be the ordinary accepted standards?

THE WITNESS: Yes, I can.

THE COURT: And what is that?

* * *

THE WITNESS: Based on what I have seen from the protocol supplied by the Houston Police Department, there are areas that I need clarification on.

THE COURT: Okay. Your answer is no. Your request for a hearing is overruled. Thank you ma’am.

The defense pressed on for a hearing on the second two Kelly prongs:

This is our request. That it requires a hearing outside the presence of the jury to determine whether the technique protocols used by the HPD lab in this case meet those standards, whether those standards that they used are subject to peer review and is there a problem with those.

. . . [W]e’re asking for the right to make this record regarding just specifically what we would have been asking you before we even get to the jury.

* * *

THE COURT: You have told me so far that you know of no specific situation that exists in terms of the collection, the retention and so on and so forth of the evidence about which you have a dispute.

You simply want to satisfy yourself that the methods used were, in fact, consistent with the scientific standards.

The court went on to explain that its reason for cutting off Dr. Johnson was her response that she needed some clarification about HPD procedures: “[w]ell, that tells me that there is no dispute. She doesn’t know enough and that this is investigatory as opposed to anything else.”

Counsel repeated that the defense had a specific dispute with HPD protocols and relayed specific concerns that Dr. Johnson expressed to him concerning the “match windows” used by HPD. Based on this discussion, the court concluded:

I’ll let you present whatever rebuttal evidence, impeachment evidence you care to the jury. In terms of an out-of-the-presence hearing, I’m not going to do it. . . . But please don’t misinterpret what I’m saying to be a refusal or an imposition of you being permitted to present this information to the jury.

I’m specifically saying I’m not doing that.

The trial court’s ruling deviated from Kelly and its progeny in two respects. First, it presumed that DNA evidence is admissible in general and shifted to the defense the burden of demonstrating that the State’s evidence was not reliable. Further, it refused to permit a hearing outside the jury’s presence to determine whether the evidence was admissible. Thus, the court erred in failing to determine the admissibility under Rule 702 by interpreting the second and third Kelly requirements as issues that the jury could consider in deciding what weight to give the State’s evidence.

Although we agree with appellant that the court abandoned its “gatekeeping” function with respect to the last two Kelly requirements, we hold that in this case the error was harmless because the State’s DNA evidence was in fact reliable. Appellant urges that on cross-examination trial counsel undermined the State’s DNA evidence by revealing that the State did not use “reagent blanks to eliminate the possibility of cross- contamination” and that the allele, which was isolated from the metal bar and attributed to appellant, was weak and could not be ruled out as a “an ‘artifact’ or nonspecific showing of DNA testing.”⁷ Upon an independent review of the record, we find that the State’s DNA evidence was reliable and thus admissible.⁸

Mary Childs Henry, a DNA analyst with the Houston Police Department performed the DNA extractions for the PCR and RFLP tests. She testified that she followed the protocol designated by the FBI at that time, which did not require reagent blanks. Although the protocol was later upgraded to include reagent blanks, use of these controls would have violated FBI protocol at the time she performed the tests. She explained that the protocol directed her to work only on one sample at a time to eliminate the possibility of cross-contamination.

Joseph Chu, a serologist with the HPD crime lab, testified that, from the mixture of DNA isolated from the metal bar, he identified an antigen that matched appellant’s DNA. He explained that the antigen was weak and could have been only a trace, but maintained that appellant could not be excluded as a source of DNA found on the bar. In Chu’s opinion, no cross-contamination or cross-hybridization occurred during the testing process.

Nothing in the trial testimony suggests that HPD serologists used invalid protocols or deviated from the proper protocols. Therefore, we find the State’s DNA evidence reliable and admissible under Kelly. Because the court’s failure to hold a hearing outside the jury’s presence did not result in the admission of unreliable evidence, the error did not have “a substantial and injurious effect or influence in determining the jury’s verdict.” King v. State, 953 S.W.2d 266, 271 (Tex. Crim. App. 1997) (announcing the test for determining whether trial error affects a substantial right under Rule of Appellate Procedure 44.2(b) (quoting Kotteakos v. United States, 328 U.S. 750, 776 (1946))). Accordingly, Rule 103(a)⁹ compels us to disregard this error. Appellant’s third point of error is overruled.

Appellant contends in his fourth point of error that the trial court should have excluded the State’s DNA evidence because the State violated a discovery order which required it to disclose the results of all scientific testing by March 5, 1998. Appellant complains that the State failed to timely turn over new statistical calculations of the frequency that the DNA profile of the blood on the two towels occurs in the population. Instead, the State informed appellant of the recalculations on the evening of March 10, the day before the State began presentation of its DNA evidence. Upon counsel’s objection to the tardy disclosure, the trial court asked how much time the defense desired to look over the new calculations. Dr. Johnson, the defense expert, replied, “if they give me the figures and the type results, . . . I don’t have that material — I would need a[n] hour or two hours to look at it.” The court granted appellant this additional time to review the information before the State presented its statistical evidence.

During the afternoon session of March 11^th, Mary Childs Henry testified that she performed RFLP tests on DNA isolated from a beige and a red towel. The DNA profile of the blood on the towels matched appellant. The original frequency calculation that Henry performed indicated that one in every 5.8 million persons possesses this DNA profile. Henry explained that she used four probes to achieve this original calculation; however, two of the probes violated FBI protocol for inclusion in the frequency analysis. When she recalculated the statistics using only the two compliant probes, she arrived at a significantly different result of one out of 224 persons.¹⁰

We have held that evidence willfully withheld from disclosure under a discovery order should be excluded from the trial. See Hollowell v. State, 571 S.W.2d 179, 180 (Tex. Crim. App. 1978). There is no indication in the record, however, whether the HPD serologist’s failure to recalculate the frequency statistics earlier was done willfully or was merely an oversight. In any event, we have distinguished the harsh sanction imposed by Hollowell, where the State revealed the requested information in time for the defendant to use it on cross examination of the State’s witness. See Smith v. State, 779 S.W.2d 417, 431 (Tex. Crim. App. 1989)(holding that State’s violation of a discovery order to disclose before trial any leniency arrangements with witnesses did not deny defendant a fair trial because the State revealed on direct examination the deal it made with a witness).

Here, appellant does not explain how he was harmed by the State’s failure to perform the recalculations earlier. He was afforded ample time for his expert to review the recalculations, and counsel was able to cross-examine the State’s witnesses with the new results. Because appellant was in no way prejudiced by the State’s failure to comply with the discovery order, we overrule appellant’s fourth point of error.

Appellant’s fifth, sixth, and seventh points of error involve alleged improper jury argument by the State. Proper jury argument includes four areas: (1) summation of the evidence presented at trial, (2) reasonable deduction drawn from that evidence, (3) answer to the opposing counsel’s argument, or (4) a plea for law enforcement. McFarland v. State, 845 S.W.2d 824, 844 (Tex. Crim. App. 1992), cert. denied, 508 U.S. 963 (1993). To constitute reversible error, the argument must be manifestly improper or inject new, harmful facts into the case. Gaddis v. State, 753 S.W.2d 396 (Tex. Crim. App. 1988); Everett v. State, 707 S.W.2d 638, 640 (Tex. Crim. App. 1986). Here, we find that each of the challenged passages fits within one or more of the categories enumerated above.

In his fifth point of error, appellant contends that during final rebuttal argument the State impermissibly shifted the burden to the defense. Specifically, appellant complains of the following passage in which the State argued that no defense experts challenged its DNA evidence.

We had two samples from the door. We had a scraping and a swab, one of those is B type . . . And don’t you know if there was anything wrong — I mean they’re accusing Mr. Bolding of lying about his results is what they’re doing here. Don’t you know if there was anything wrong or improper about those results — and understand, they don’t have to do anything. They can sit there like potted plants. But don’t you know if there was something wrong, they would have had one of those experts on the stand to tell you that that evidence wasn’t B type. And you can make an inference from that.

Then you come to the DNA. And it follows here as well. I mean, they haven’t had just one, but they’ve had two experts looking at our DNA evidence. They’ve had one of them sitting in the courtroom listening to the testimony. Don’t you know if there was anything wrong with the results of those samples — [¹¹]

Our review of the record reveals that the prosecutor’s comments were rebuttal to defense counsel’s argument attacking the State’s scientific evidence. Arguing that the State did not produce reliable evidence that the blood on the wall matched appellant, counsel pointed out that James Bolding, the State’s serology expert, waited ten years to conduct the tests revealing type-B blood on the wall. Counsel also implied that the defense DNA expert, who had been sitting at counsel table during the State’s presentation of its case, did not approve of the State’s DNA evidence; particularly, the dramatic change in the frequency statistics when the two noncompliant probes were eliminated. Counsel continued with arguments attacking the reliability of the PCR testing done on the metal bar, the State’s use of improper protocol, and the methods used to extract the DNA.

Defending the State’s scientific evidence, the prosecutor retorted that the defense would have called its expert to the stand if it had seriously disputed the State’s evidence. Thus, the prosecutor’s argument was merely a response to the defense’s assertions that the State’s scientific evidence was lacking. See McFarland v. State, 989 S.W.2d 749, 751 (Tex. Crim. App. 1999). In this context, the challenged comments did not impermissibly shift the burden of proof to the defense, particularly given the prosecutor’s repeated reminders to the jury that the defense was not required to present any evidence.

Moreover, the prosecutor’s argument did not amount to a prohibited reference to appellant’s failure to testify. We have held that the prosecutor may comment on the defendant’s failure to produce witnesses and evidence so long as the remark does not fault the defendant for exercising his right not to testify. Patrick v. State, 906 S.W.2d 481, 491 (Tex. Crim. App. 1995), cert. denied, 517 U.S. 1106 (1996); Livingston v. State, 739 S.W.2d 311, 338 (Tex. Crim. App. 1987), cert. denied, 487 U.S. 1210 (1988). Here, reference to the defense’s failure to produce expert testimony was not improper. Appellant’s fifth point of error is overruled.

Appellant next asserts that during punishment-phase argument the prosecutor improperly commented on facts outside the record, as follows:

But here at punishment it is different. You get to hear more. You get to know the real Derrick Jackson. And you get to take into consideration his victims, the impact on the victims’ family and society. You get to look at all that. Now a life sentence to you — you and me would be a lot of punishment. I am going to suggest to you . . .

Defense counsel interrupted with an objection that the prosecutor was interjecting his own opinion into the case. The court sustained the objection, instructed the jury to disregard, and ordered the prosecutor to rephrase. The prosecutor continued, “Let’s put it this way: A life sentence would be punishment for a lot of people, but not to Derrick Jackson.” Appellant contends that the trial court should have granted his motion for mistrial because the prosecutor’s comments about a life sentence “suggested that he has information about the Institutional Division of the Texas Department of Criminal Justice (“TDCJ”) to which the jury was not privy.”¹² The record does not support this inference.

A prosecutor may not convey to the jury during argument that he possesses specialized knowledge or expertise about a contested factual issue in the case. Such comments pose a danger of influencing the jury’s opinion in deciding the issue. Johnson v. State, 698 S.W.2d 154, 167 (Tex. Crim. App. 1985), cert. denied, 479 U.S. 870 (1986); Boyd v. State, 643 S.W.2d 700 (Tex. Crim. App.1983); Maupin v. State, 930 S.W.2d 267, 270 (Tex. App.—Fort Worth 1996, pet. ref’d). In evaluating whether the prosecutor interjected his personal opinion into the argument, we must consider the challenged remark in the context in which it appears. Gaddis, 753 S.W.2d at 398.

When we evaluate the prosecutor’s statements in the appropriate context, it is evident that his remark concerning a life sentence referred to appellant’s apparent contentment with the prison environment. The prosecutor began his argument, “Now, this phase of trial, it’s the punishment phase. It’s not the what’s best for Derrick Jackson phase.” He went on to summarize appellant’s punishment-phase evidence that appellant performed well in prison, reacted positively to the structured setting, and could accomplish many of his goals there, such as earning a GED. The crux of the argument, therefore, was that prison would not serve to punish appellant. Within this context, it is clear that the prosecutor’s statements comparing the punitive effect of a life sentence on appellant to a life sentence on people in general—such as the jurors—was based on appellant’s proffered evidence, as opposed to some specialized knowledge harbored by the prosecution. Accordingly, the prosecutor’s comments qualify as a permissible plea for law enforcement. Appellant’s sixth point of error is overruled.

In his seventh point of error, appellant contends that the prosecutor impermissibly argued facts outside the record and thus misconstrued the testimony of the defense’s mental health expert. The challenged portion is as follows:

And Dr. Wheeler is tremendously concerned about that kind of individual if they ever got out of the penitentiary. She’s testified in 12 capital murder cases for the Defense. And the person that did this scares the hell out of her. And she told you under oath that she would never want a person capable of doing all that on the streets of Harris County, Texas again.

Defense counsel objected that the prosecutor was misstating Dr. Wheeler’s testimony. The trial court sustained the objection and instructed the jury to disregard the comment, but denied the defense’s motion for mistrial. The prosecutor continued, “She told you that in her opinion that kind of person would be capable of doing all of this again.” The trial court overruled appellant’s objection that this comment was outside the record. Appellant asserts that the trial court should have granted his motion for mistrial because the prosecutor impermissibly argued outside the record to undermine the defense’s expert witness.

Contrary to appellant’s assertion, however, the record reveals that the prosecutor’s argument was properly based on the evidence. During the punishment phase, the prosecutor questioned Dr. Wheeler:

Doctor, in your professional opinion, you wouldn’t want someone that’s capable of committing a capital murder where two people are killed like you see in those pictures, stealing a purse from a lady, during an aggravated robbery, they shoot at somebody and then doing three more aggravated robberies [. . .] a week later within an hour and a half, you wouldn’t want somebody like that back on the streets of Harris County, Texas would you?

Dr. Wheeler replied, “No, I would not.” Wheeler also agreed on cross-examination that she would be scared of a person who had just committed these various crimes. Although the prosecutor’s characterization of Wheeler’s testimony is somewhat overstated, we find that this argument qualifies as a summation of the evidence, and thus is not improper. Appellant’s seventh point of error is overruled.¹³

In his eighth point of error, appellant complains that the trial court abused its discretion in failing to grant a motion for mistrial when the jury appeared unable to agree on sentencing. At the end of the punishment phase, the jury deliberated for about five hours and sent out a note indicating that they were deadlocked. After the court instructed them to continue their deliberations, they deliberated for another two hours at which time counsel for appellant moved for a mistrial. The court overruled the motion, and the jury recessed for the night. On the following day, the jury had deliberated for an additional four hours when they sent out a note stating, “What happens if we are 11 to 1 on special issue #3?” The court again instructed them to continue deliberating.¹⁴ The jury returned a unanimous verdict 28 minutes later.

Article 37.0711, § 3(g) compels the trial court to enter a life sentence if the jury is unable to answer any special issue. See Montoya v. State, 810 S.W.2d 160, 166 (Tex. Crim. App. 1989), cert. denied, 502 U.S. 961 (1991). The court may exercise its discretion to discharge the jury when “it has been kept together for such a time as to render it altogether improbable that it can agree.” Art. 36.31. Reversal is mandated only if the record reveals that the trial court abused its discretion in holding the jury for deliberations. Montoya, 810 S.W.2d at 166. In assessing abuse of discretion, we consider the length of the trial and volume of evidence submitted to the jury. Howard v. State, 941 S.W.2d 102, 121 (Tex. Crim. App. 1996). For instance, in Andrade v. State, 700 S.W.2d 585, 589 (Tex. Crim. App. 1985), cert. denied, 475 U.S. 1112 (1986), the trial court instructed the jurors to continue deliberations when, after four-and-a-half hours, they indicated that they were deadlocked. The jury persisted with its deliberations for another eight hours before returning an unanimous verdict. Id. Given the nature of the capital case involved, this Court determined that the judge did not abuse his discretion in permitting the jury to deliberate for twelve hours. Id.

Likewise, in this case, the jury was charged with the task of answering questions that determine whether appellant would receive a sentence of life or death. When considering appellant’s motion for mistrial, the trial court explained that the guilt/innocence and punishment phases of trial had produced three-and-a-half days worth of evidence. Over this period of time, the jury heard testimony from 27 witnesses for the State and defense. During the punishment phase, the State produced evidence of the numerous violent crimes appellant committed. Appellant countered with testimony from an expert concerning his psychological profile, from his parents that he was generally well behaved, and from his TDCJ barber school teacher that he had adapted well to prison life, was a good student, and had successfully completed 1400 hours of course work. The jury had to sort through this evidence and evidence presented at the guilt/innocence phase to determine whether the mitigating evidence compelled a sentence of life rather than death. See Art. 37.0711 § 3(e). We cannot say, given these circumstances, that the trial court abused its discretion by overruling appellant’s motion for mistrial seven hours into deliberations. Appellant’s eighth point of error is overruled.

In his final point of error, appellant argues that the trial court should have granted his challenge to the constitutionality of the Texas death penalty scheme based on the “12- 10” rule embodied in Article 37.0711, which prohibits the jury from learning that a hung jury on punishment results in a life sentence. See Art. 37.0711 §§ 3(d), (f), and (i). He points out that in this case the jury was deadlocked twice during deliberations and had inquired as to the effect of a hung jury. Appellant is correct in his concession that this Court has repeatedly rejected such an attack. See, e.g., Jackson v. State, 992 S.W.2d 469, 481 (Tex. Crim. App. 1999); Cantu v. State, 939 S.W.2d 627, 644 (Tex. Crim. App. 1997), cert. denied, 522 U.S. 994 (1997); Green v. State, 912 S.W.2d 189, 194 (Tex. Crim. App. 1995), cert. denied, 518 U.S. 1021 (1996). We overrule appellant’s ninth point of error.

Finding no reversible error, we affirm the judgment of the trial court.

Price, J.

Date Delivered: May 17, 2000

Publish
¹ Unless otherwise indicated, all future references to Articles refer to Code of Criminal Procedure.
² Lindgroff was deceased at the time of trial.
³ Henderson’s car was involved in a burglary of a Montgomery Ward two to three days later. Police engaged in a high speed chase with the perpetrators, who wrecked the car and fled before police could catch them.
⁴ In keeping with this argument, appellant must have entered the apartment after all of this activity, left a bloody fingerprint on the bedroom door, cut himself, wiped his wound on the bathroom towels, locked the door, and left.
⁵ Two types of DNA analysis were utilized in this case—PCR and RFLP. “PCR,” stands for “polymerase chain reaction,” a method of multiplying DNA so that a profile can be obtained from a very small sample. “RFLP” stands for “restriction fragment length polymorphism,”a method of DNA analysis that produces a more discriminating profile but requires a larger DNA sample. See, e.g., Campbell v. State, 910 S.W.2d 475, 479 n.6 (Tex. Crim. App. 1995).
⁶ The Texas Rules of Evidence (formed from the combined civil and criminal rules of evidence) became effective on March 1, 1998.
⁷ Appellant does not complain now of the match windows, protocol, or statistical analysis used. At trial, the court afforded Dr. Johnson, the defense expert, time to review the DNA type strips and statistical figures that were in the State’s possession. Presumably, Johnson was satisfied with these because appellant did not call her to testify before the jury.
⁸ We are authorized to conduct a de novo review of this mixed question of law and fact because the admissibility issue does not turn on the credibility and demeanor of the State’s witnesses. See Guzman v. State, 955 S.W.2d 85, 87 (Tex. Crim. App. 1997).
⁹ Rule 103(a) provides that“[e]rror may not be predicated upon a ruling which admits or excludes evidence unless a substantial right of the party is affected.” Tex. R. Evid. 103(a).
¹⁰ Joseph Chu testified later that the statistical calculation based on the PCR test conducted on this same blood source indicated that the DNA would occur with a frequency of “one out of 7.2 million people among the black population.” Apparently no recalculations were performed on the genetic profile resulting from the PCR analysis.
¹¹ Defense counsel interrupted with an objection that the State was shifting the burden of proof to the defense, and the court overruled the objection.
¹² Appellant notes that the prosecution possessed information that the jury did not—that appellant would have become parole eligible in fifteen years if given a life sentence. Appellant did not raise this parole issue to the trial court; therefore, he has waived such an argument for appeal. See Johnson v. State, 803 S.W.2d 272, 292 (Tex. Crim. App. 1990), cert. denied, 501 U.S. 1259 (1991).
¹³ Appellant concedes that this error alone would not constitute reversible error but contends that the statements challenged in points of error five, six, and seven—when considered in combination—deprived him of a fair trial. Finding no error in the prosecutor’s argument, we reject appellant’s contention.
¹⁴ Appellant did not reassert his motion for mistrial at this time.