Ex Parte Mowbray

Annotate this Case

943 S.W.2d 461 (1996)

Ex parte Freda S. MOWBRAY a/k/a Susie Mowbray.

No. 72624.

Court of Criminal Appeals of Texas, En Banc.

December 18, 1996.

Rehearing Denied February 12, 1997.

Robert Ford, Fort Worth, for appellant.

John A. Olson, Robert Moore, Asst. District Attorneys, Brownsville, Matthew Paul, State's Atty., Austin, for State.

Before the court en banc.

OPINION

BAIRD, Judge.

Applicant was convicted of murder and sentenced to confinement for Life and a $10,000.00 fine. The Court of Appeals affirmed. Mowbray v. State, 788 S.W.2d 658 (Tex. App.Corpus Christi 1990). Her habeas application contends: she is actually innocent of murder; the State's expert knowingly gave false and misleading testimony; the State knowingly used false testimony; and, she suffered ineffective assistance of counsel at trial. The habeas judge entered findings of fact and conclusions of law, recommending that we grant a new trial. We will grant relief.

I.

The evidence at trial was summarized by the Court of Appeals as follows:

The deceased was shot in bed at night. The only occupants of the room in which the shooting occurred were the deceased and [applicant]. *462 The defense theory was that [applicant] and the deceased were lying in bed with a pillow barrier between them when [applicant] saw the deceased's elbow point upward. When she reached to touch it, the gun went off. [Applicant] made a taped statement about the shooting, and the tape was admitted into evidence. Witnesses to the [applicant's] statements recalled that [applicant] indicated that she had used her left hand to reach toward the deceased. The State, however, introduced a crime lab supervisor's analysis of [applicant's] nightgown showing traces of lead or gunshot residue on the lower right sleeve. That witness, Steve Robertson, conducted tests with the gun found at the scene and opined that the residue was consistent with someone firing that gun. Estella Mauricio, who was dispatched to the Mowbray residence just after the shooting, testified that she found the deceased, still alive and shot through the head, lying on his left side and covered all the way up to his shoulder. The bullet had entered the right side of his head, exited to the left, and wounded his left hand, which was under his head with a pillow between his head and left hand. The right hand was lying across his chest under the covers. There was no blood or brain matter on the right hand and she did not ever see his hand being washed at home or at the hospital. Emergency technician Cavazos also recalled that the victim was completely covered, with only the right side of his face and the top of his head showing, when he arrived. Mauricio testified that he was positioned like a sleeping person, and the deceased's first cousin, Scott Mowbray, testified that the deceased slept in the position in which he was found. Dr. Dahm, who conducted the autopsy, stated that if the deceased had shot himself, his right hand would have been covered with blood and brain matter. He found no such blood or brain matter on the deceased's right fingers, hand, or forearm. Dahm testified it would have been impossible for the deceased to have shot himself and the hand to be clean, and concluded that the death was a murder.....

Id., 788 S.W.2d at 662-663.

Additionally, two blood spatter experts testified. Sergeant Dusty Hesskew, of the Austin Police Department testified on behalf of the State and Captain Tom Bevel of the Oklahoma City Police Department testified on behalf of applicant. Generally, blood spatter experts inspect the physical evidence to determine the injuries suffered and their location with respect to the other physical evidence. In the instant case, both experts examined applicant's nightgown for "high velocity impact [blood] staining" which commonly occurs within a short distance from a contact gunshot wound. Hesskew testified that he identified and measured, through "luminol testing," high velocity impact bloodstains on applicant's nightgown, which were invisible to the naked eye. Hesskew concluded the cause of death in the instant case was probably homicide. Bevel testified that his examination of the physical evidence led him to conclude the deceased could have died in the manner in which applicant testified, i.e., suicide.

II.

Pursuant to our order, the habeas judge conducted a hearing on the instant application. At that hearing, a third blood spatter expert, Herbert Leon MacDonell, the director of an independent forensic laboratory in Corning, New York, testified. MacDonell is viewed as the pre-eminent authority on the science of blood spatters.[1] MacDonell was retained to review the photographs and physical evidence in the instant case by the Cameron County District Attorney's office approximately seven months prior to trial. MacDonell's examination of applicant's nightgown revealed no blood stains either visible to the naked eye or under a microscope. MacDonell concluded: "[I]t is very unlikely that [applicant's nightgown] was in close proximity to the victim's gunshot wound at *463 the time of his shooting, or it was protected from spatter in some manner if [it] were." After reviewing the crime scene, the physical evidence and the photographs, MacDonell's expert opinion was that it was more probable than not that the deceased died from a suicide rather than a homicide. At the request of the State, MacDonell prepared and mailed to the Cameron County District Attorney a written report of his findings approximately two weeks before trial.

MacDonell took issue with Hesskew's use of luminol to measure blood spattering. Luminol is a substance which can react with blood that is invisible to the naked eye. However, luminol testing is not accepted as a positive test for blood. Luminol testing is merely presumptive because luminol reacts with substances other than blood. In MacDonell's opinion, the luminescence from a luminol reaction cannot be accurately measured. He stated:

I think it would truly be an exercise to futility. I don't think you can put any reliability on itI certainly wouldn'tand I've seen luminol sprayed many times. I've never heard of anyone trying to measure it, count it, other than saying there appears to be a dozen or more.... You could do it, but the validity of your conclusion would be highly suspect in my opinion.

In MacDonell's view, Hesskew did not understand the chemistry behind luminol testing.[2]

Hesskew testified he was retained by the Cameron County District Attorney's office as a blood spatter expert. He observed applicant's nightgown at the Department of Public Safety laboratory prior to the time it was shown to MacDonell. Hesskew was present when the nightgown was treated with luminol. He counted forty-eight small stain areas around the stomach and chest of the nightgown which appeared consistent with high-velocity stains. Hesskew further donned a similar nightgown and fired test shots into a CPR dummy's head filled with blood in an attempt to duplicate the staining he observed through the luminol testing. Although Hesskew could not remember how he was able to duplicate the blood staining, in his expert opinion, applicant, wearing her nightgown, could not have been laying beside her husband at the time of his death. Thus, Hesskew's testimony contradicted applicant's defensive theory.

Hesskew admitted that his testimony at trial included several assumptions which involved more than his own test results. Most important of these assumptions was that someone tested the invisible stains and determined them to be human blood. At the hearing on the instant habeas application, Hesskew conceded his trial testimony was scientifically invalid because no such confirmation was ever made. In other words, Hesskew conceded that his ultimate opinion that the victim died as a result of a homicide, and that applicant's statements were impossible, had no scientific basis.

Bevel testified that it is impossible to measure high velocity impact blood spatter in the manner utilized by Hesskew. Bevel, like Hesskew, only performed presumptive tests on applicant's nightgown. Bevel did not conduct any confirmation tests because he was informed by Hesskew that the Department of Public Safety laboratory confirmed human blood on applicant's nightgown. Because his trial testimony was based upon this erroneous premise, Bevel stated: "with the inability to determine that ... is blood that is there, especially since we are talking about blood that is only invisible to the unaided eye, I don't think you can really say anything." Bevel believed the failure to conduct confirmation tests undermined his examination and earlier testimony, and agreed with Hesskew that their trial testimony was scientifically invalid.

Steve Robertson, a chemist in the Texas Department of Public Safety (DPS) crime laboratory, testified that he examined applicant's nightgown and was present on three different occasions when the nightgown was sprayed with luminol. The nightgown was also sprayed with three chemicals to determine *464 the presence of lead residue and treated with heat and chemicals to determine the presence of gunshot residue. His examination of the nightgown revealed very small red stains, visible to the naked eye, lead residue and a yellowish stain. Robertson conducted two confirmatory tests on the red stains to determine if they were human blood. Both tests resulted in negative results. Robertson testified that, if the stains were blood, the tests for the gunshot residue could have destroyed the protein in the blood and would cause a negative reaction. Further, the chemicals sprayed on the nightgown could have diffused or dissolved the red stains to the extent they were undetectable without a microscope.

Ed Cyganiewicz, the lead prosecutor at trial, conceded that the State's case "depended upon" the blood spatter evidence. Cyganiewicz testified that MacDonell was retained by the Cameron County District Attorney's office as an expert for the State and a subpoena was issued to secure his testimony. According to Cyganiewicz, the State did not call MacDonell to testify because of the expense of securing this testimony. However, because of the possibility of MacDonell's report being "Brady material," a copy was forwarded to applicant's trial attorney ten days to two weeks prior to trial.

Abel Toscano, a board certified criminal law specialist, had been a criminal defense attorney in Cameron County for forty-two years and was retained to represent applicant at trial. Toscano received a copy of MacDonell's report shortly before trial after the trial judge threatened Cyganiewicz with sanctions. After discussing the matter with co-counsel and Bevel, and reviewing MacDonell's report, the evidence he had gathered independently, including the statements and information he obtained from applicant, Toscano chose not to contact MacDonell because he (Toscano) did not know how MacDonell would respond and there was some chance MacDonell might change his opinion. However, Toscano did make notations in preparation of a cross-examination of MacDonell, and during trial had his assistants checking to see if MacDonell had arrived.

III.

Following the hearing the habeas judge entered findings of fact and conclusions of law. The habeas judge determined that applicant was not factually innocent under State ex rel. Holmes v. Honorable Court of Appeals for Third District, 885 S.W.2d 389 (Tex.Cr.App.1994); that, although the State's expert's testimony was scientifically invalid, it was not perjurious; and, that applicant's trial counsel was not ineffective. Nevertheless, the habeas judge determined that applicant was denied due process. Among the habeas judge's findings of fact are:

7. There was rationale for both murder and suicide.... 8. The rationale for suicide was, at least, equally persuasive.... The decedent had attempted suicide at least twice prior to his death and had shot himself in one of these previous efforts. Decedent had vowed to kill himself....[3] 11. The linchpin of the State's case was the high velocity impact spatter (HVIS) alleged found on the front of her gown. If there, [applicant] could not have been prone in the bed at the time the shot was fired, and it followed her version of events was false. 13. At [the habeas hearing] Hesskew recanted his trial testimony as to HVIS. Hesskew admitted his testimony at trial was without "scientific validity." 15.... Bevel, too, disavowed his trial testimony. 16. Without Hesskew's testimony, and the conclusions he drew, there was another equally reasonable hypothesis other than Applicants guilt: Mowbray's death was suicide or an accident. However, the matter does not rest here. The facts relating to Herbert Leon MacDonell now loom large. 17. The State's conduct in connection with MacDonell was, at best, questionable trial strategy, and, at worst, intentional deception of Applicant's counsel. *465 20.... On [MacDonell's] laboratory work sheet, dated November 18, when the State's counsel were in Corning, MacDonell made the entry: "Night gown has no pattern of high velocity impact spatter"

On a billing submitted to the State by MacDonell there is a charge of $500.00 for "18 Nov 87 Conference in Corning, New York, to examine and discuss certain physical evidence." There seems to be no question that on November 18, 1987, seven months before trial, the State's [counsel] knew MacDonell's findings and that they were exculpatory. [Emphasis in original.]

28. This Court heard and observed MacDonell as well as Hesskew and Bevel. MacDonell was far and away the more articulate, knowledgeable, and credible. If the jury had heard MacDonell, and Hesskew had testified as he now does, or not testified at all, there is a substantial probability a verdict of Not Guilty would have ensued.

Under these facts, the habeas judge determined the State violated applicant's due process right to a fair trial by suppressing evidence favorable to applicant. Indeed, the habeas judge concluded:

Applicant's counsel were not ineffective. They vigorously defended their client in a wholly professional manner. However, their effectiveness was, in large measure, thwarted by the conduct of the State.... Undoubtedly, if [Applicant's] counsel had insisted that the State produce its subpoenaed witness MacDonell, and had examined him, a different verdict might have resulted. But Applicant's counsel could not have known the devastating testimony MacDonell would have given (and did give at the writ hearing). They knew nothing more than what MacDonell had said in his written report of May 13. On the other hand State's counsel early on recognized the potential lethal effect of MacDonell's testimony on their theory of the case, and beginning in November and continuing until May they engaged in a deliberate course of conduct to keep MacDonell's findings and opinions from Applicant's counsel until the last days before trial. Even then they caused Applicant's counsel to believe MacDonell would be a witness and available for cross-examination. The State had an obligation to be forthcoming when the Brady motion was heard and granted in March. Instead, it chose to suppress MacDonell's exculpatory evidence until its hand was forced by the trial judge only days before trial, and, in so doing, the State denied Applicant due process.

Thus, the habeas judge recommends that we set aside applicant's conviction and order a new trial.

IV.

Virtually every fact finding involves a credibility determination. We have repeatedly recognized that the fact finder is the exclusive judge of the credibility of the witnesses. See e.g., DuBose v. State, 915 S.W.2d 493 (Tex.Cr.App.1996); Joseph v. State, 897 S.W.2d 374 (Tex.Cr.App.1995); Muniz v. State, 851 S.W.2d 238 (Tex.Cr.App. 1993); Bonham v. State, 680 S.W.2d 815, 819 (Tex.Cr.App.1984); Mills v. State, 508 S.W.2d 823, 826 (Tex.Cr.App.1974); Esquivel v. State, 506 S.W.2d 613 (Tex.Cr.App.1974); and, West v. State, 489 S.W.2d 597 (Tex.Cr. App.1973). In habeas hearings, the judge determines the credibility of the witnesses and if the habeas judge's findings of fact are supported by the record, they should be accepted by this Court. Ex parte Turner, 545 S.W.2d 470, 473 (Tex.Cr.App.1977). See also, Ex parte Adams, 768 S.W.2d 281, 288 (Tex. Cr.App.1989).

In Ex parte Moore, 136 Tex.Crim. 427, 126 S.W.2d 27 (1939), we stated:

... Where the ruling of the trial judge depends upon the existence or non-existence of a certain fact and testimony pro and con is introduced thereon and the evidence is conflicting it becomes the duty of the trial judge to determine the issue, and unless it appears to this court that his finding was without support in the evidence, and that he had committed an error in his judgment thereon, we would not *466 interfere with his findings thereon. Glenn v. State, 89 Tex.Cr.R. 13, 229 S.W. 521.

Id., 126 S.W.2d at 28.

The State has an affirmative duty to disclose favorable evidence under the Due Process Clause. Brady v. Maryland, 373 U.S. 83, 87, 83 S. Ct. 1194, 1196-1197, 10 L. Ed. 2d 215 (1963). However, in order to warrant a new trial, the favorable evidence must also be material. Ex parte Castellano, 863 S.W.2d 476, 485 (Tex.Cr.App.1993).[4]

In the instant case, the habeas judge determined the State knowingly suppressed MacDonell's report which supported applicant's theory of suicide. This was obviously favorable evidence. Additionally, the habeas judge found that if the favorable evidence had been considered by the jury, an acquittal would have resulted. Clearly this finding meets the materiality requirement. Id., 863 S.W.2d at 486.

The habeas judge's factual determinations are supported by the record and, therefore, will be accepted by this Court. Turner, 545 S.W.2d at 473. Consequently, we hold applicant's due process rights were violated, and applicant is entitled to relief. Accordingly, applicant's conviction is set aside and it is the order of this Court that applicant be returned to the Sheriff of Cameron County to answer the indictment.

Relief granted.

WHITE, J., not participating.

McCORMICK, Presiding Judge, dissenting.

I must respectfully dissent. In this case, the majority grants relief on this eight-year-old conviction by holding applicant's due process rights under the Fourteenth Amendment to the United States Constitution were violated when the State "suppressed" MacDonnell's report by furnishing it to applicant about two weeks before her 1988 trial began! Just to state the Court's holding is to refute it. The Court's holding as to this aspect of the case[1] is absurd and represents another inane exercise in judicial activismi.e., finding so called "rights" in the United States Constitution that just do not exist.

I cannot add much more to Judge Keller's dissenting opinion which I also join. MacDonnell's report does not in any way impeach the most important evidence in this case establishing applicant's guiltthe absence of blood and brain matter on the victim's right hand, the position of the victim's body, and the presence of gunshot residue on the right sleeve of applicant's nightgown all of which were inconsistent with her defensive theory. See Mowbray v. State, 788 S.W.2d 658, 663 (Tex.App.Corpus Christi 1990, pet ref'd), cert. denied, 498 U.S. 1101, 111 S. Ct. 999, 112 L. Ed. 2d 1082 (1991). The evidence at applicant's 1988 trial amply supports a finding that applicant shot and killed the victim while he was asleep in bed, and MacDonnell's report fails to cast any serious doubt as to the reliability of this finding. I would hold applicant fails to present any cognizable claim for habeas corpus relief. Cf. Herrera v. Collins, 506 U.S. 390, 113 S. Ct. 853, 122 L. Ed. 2d 203 (1993).

In addition, the basis of applicant's claims for relief in this proceeding has been known to applicant since about two weeks before her 1988 trial. However, she failed to raise these claims at the motion for new trial hearing following her 1988 conviction or during the direct appeal of her 1988 conviction. See Mowbray, 788 S.W.2d at 662-72, 669 (applicant filed motion for new trial claiming the State failed to produce the T-shirt the victim wore when shot).[2] Applicant has delayed *467 approximately seven years before bringing her claims for relief in this proceeding. Under these circumstances, the State's and society's legitimate and valid interest in the finality of this obviously guilty applicant's conviction should outweigh granting applicant relief on any due process violation that allegedly occurred in connection with her 1988 trial. Someone should be asking whether the State can get a fair trial after such a long passage of time, and whether applicant's delay in asserting her claims for relief in this proceeding constitutes a waiver of those claims.

I respectfully dissent.

MANSFIELD and KELLER, JJ., join this dissent.

KELLER, Judge, dissenting.

Applicant makes three claims in her effort to gain relief from her conviction. The trial court found that applicant was not entitled to relief on the basis of any of her three claims, but recommended relief anyway, determining that the State's suppression of favorable evidence denied applicant due process.

The majority's decision that applicant's conviction should be set aside rests upon the determination by the trial court that the State knowingly suppressed MacDonnell's report, and that if the jury had considered the report, there is a substantial probability applicant would have been acquitted.[1]

The record does not support the finding of the trial court that the State engaged in a deliberate course of conduct to keep MacDonnell's findings and opinions from defense counsel until the last days before trial. The prosecutor wrote twice to MacDonnellin March and in Aprilto request him to forward his written findings. No report was sent. The report in question was not prepared by MacDonnell until May 13, 1988. The State then turned the report over to the defense immediately, ten days before the trial began on May 23, 1988.

Neither does the record support the finding that the State misled defense counsel by representing that it would produce MacDonnell at trial. Although the State retained MacDonnell and issued a subpoena for him, this was not shown to be a subterfuge to mislead counsel: the State did not get the final report until May of 1988 either.

But even if the record did support the trial court's finding that the State suppressed the report,[2] applicant would not be entitled to relief. Her attorney did know of the report before trial; he and his client made a tactical decision not to follow up on the report.

The full reasons for this decision are unknown to us for the sole reason that Applicant would not allow trial counsel to testify to all the reasons he decided not to use MacDonnell or the report.[3] Regarding the reasons Toscano did not use the report or MacDonnell, Toscano stated:

There were a number of other factors. And if the Court will allow me to do so byI would have to reveal some confidential information that I don't think is proper unless the Court orders me ... I feel that I should be released of this confidentiality so as to bring forth before this tribunal the facts that wethat wethat you should know, what the bases were why I did not call Mr.call or talk to him.

(Ellipses added.) The trial court responded that applicant had to waive the privilege. Applicant consulted with her attorney off the record, with the result that the privilege was not waived. Because applicant purposely kept from the trial court the full reasons that she and her attorney decided not to use the report or MacDonnell, we must assume that *468 the decision was sound and that the defense did not consider the report to be favorable.[4] And if we make that assumption, there is no basis for concluding that the "suppression" of the report denied applicant due process or, indeed, had any deleterious effect on the trial at all.

MacDonnell's report was not the only matter in dispute at the writ hearing. Testimony at the hearing established that witnesses erroneously believed that a confirmatory test had been done on the nightgown after presumptive tests indicated the presence of blood. In fact, confirmatory tests had been done with negative results. The negative results did not necessarily indicate a lack of blood, but established only that if there was blood, the amount was too small to react or it had been degraded by previous forensic tests.[5]

Importantly, Hesskew's testimony that there was high velocity impact spatter was entirely unaffected by the fact that there was no positive confirmation that the substance on the gown was blood. That is, the pattern of the stain was caused by the atomization of the substancewhatever it wasproduced when the surface tension of the substance was overcome by very high energy from, for example, a gunshot. There was testimony that high velocity impact spatter is associated ninety-nine times out of one hundred with gunshots, known otherwise to occur in traffic fatalities when a car was traveling at a high rate of speed. The victim's death was caused by a high energy contact gunshot to the head.

Testing with luminol, with which the presumptive test was done, is relatively specific. That is, there is a limited number of substances with which it will react. There was testimony that luminol will react to blood, copper, iron, nickel, and bleach. There was also testimony that had the stain on the nightgown been a bleach stain it would not have been likely to produce the pattern that was on the gown.

The trial court's finding that Hesskew recanted his trial testimony as to HVIS is thus unsupported by the record except to the extent that Hesskew believed a confirmatory test for the nature of the substance on the gown had indicated blood. Likewise, the record does not support the finding that Hesskew disavowed virtually the whole of his trial testimony. Ultimately, the opinions he expressed at trial varied from his opinions at the writ hearing only in that he conceded that he did not prove to a scientific certainty that there was blood on the nightgown.

There was testimony at trial and at the writ hearing that red stains were observed on the nightgown, both microscopically and with the naked eye. There was evidence that the victim was found in his normal sleeping position, with his left hand under his head and a pillow between his knees, and that such a position would not be usual for a man who wished to commit suicide. There was no blood and brain matter on the victim's hand, as there should have been if he had held the gun.[6] There was testimony that the pattern of the stain on the nightgown was caused by high velocity impact spatter.

The trial court concluded that applicant failed to prove her claims of "actual innocence," ineffective assistance of counsel, or perjured testimony. These findings are supported by the record. The court's finding that HVIS was the linchpin of the State's case is not supported by the record. Even conceding that it was important, doubt was cast upon only one aspect of the HVIS testimony. Something produced a stain on the gown in a pattern that is associated almost *469 exclusively with gunshots. Under the circumstances of this case, the lack of scientific certainty as to what that substance was is not sufficient to establish a denial of due process.

Believing that applicant has failed to establish a right to relief under any of her claims or under due process, I would deny relief.

MANSFIELD, J., joins this dissent.

NOTES

[1] The State's expert, Dusty Hesskew, studied under MacDonell and testified elsewhere that MacDonell "basically invented" blood spatter analysis. Indeed, during Bevel's cross-examination, the State sought agreement that MacDonell "is THE expert or THE granddaddy of blood spatter."

[2] MacDonell further testified:

It's uncomfortable to criticize other people who are practicing in the field that I have been quite responsible for bringing to the front of the current investigative systems in the United States, if not the world, but I must say in one word that I'm disappointed. I will leave it at that.

[3] All emphasis is supplied unless otherwise indicated.

[4] In Thomas v. State, 841 S.W.2d 399, 404 (Tex. Cr.App.1992), we held:

... [T]he Due Process Clause of the Fourteenth Amendment is violated when a prosecutor fails to disclose evidence which is favorable to the accused that creates a probability sufficient to undermine the confidence in the outcome of the proceeding.

[1] The Court's opinion fails to address the issues of cognizability and procedural default. Since this proceeding is a collateral attack on applicant's conviction approximately eight years after the fact, these are important issues which deserve serious consideration.

[2] One need only examine the Court of Appeals opinion affirming applicant's 1988 conviction to see that her conviction more than satisfies due process principles. See Mowbray, 788 S.W.2d at 662-72.

[1] The majority's statement that the trial court found that "an acquittal would have resulted" is inaccurate. Op. at p. 466.

[2] Ordinarily, a due process claim relating to the suppression of favorable evidence does not apply where the defense obtained the evidence before trial. Havard v. State, 800 S.W.2d 195, 204-205 (Tex.Crim.App.1989)(opinion on original submission).

[3] Apparently, no one recognized that applicant had waived the attorney/client privilege when she alleged ineffective assistance of counsel. Toscano should have been allowed to testify as to the reasons for his decision regarding MacDonnell.

[4] At least one of MacDonnell's conclusions contradicted applicant's statement to the police and her defensive theory: MacDonnell concluded that it was unlikely that the gown was near the wound at the time of the fatal blast. Applicant had admitted, however, that she was wearing the gown and was close to the victim.

[5] There was, in fact, a considerable amount of testimony that luminol tests and tests for gunshot residue could degrade the sample to the extent that it would not be possible to get a positive confirmation of the presence of blood.

[6] The trial court found that the absence of blood and brain matter was a wash because there was no blood and brain matter on applicant's arm either. The finding that the absence of such matter was a wash is unsupported by the record; there was testimony that applicant's arm could have been shielded by something like a towel.