Genene Jones: The Test on Trial

Case No. 02

A pediatric nurse was convicted in 1984 on a toxicology test almost no one else had run. The appeal that upheld it argued for leaving that standard behind, as Texas did six years later.

The Victims

Chelsea Ann McClellan was born on June 16, 1981, about four weeks premature, and was 15 months old when she died on Sept. 17, 1982 (Texas Monthly). Her mother, Petti McClellan, called her strong-willed. Petti became a nurse herself and spent decades opposing Jones’s release and pressing for new charges. She died in June 2019, months before the case she pushed for ended in a guilty plea (Texas Monthly).

Rolando Santos was four weeks old when he was admitted to the county hospital in San Antonio with pneumonia in late December 1981 (Texas Monthly). On Jan. 9, 1982, Jones injected him with heparin, a blood thinner (TDCJ). He survived. In 2013 he spoke out against her planned release.

Joshua Sawyer was 11 months old and in the hospital for smoke inhalation after a fire at his family’s rental home. His mother was 20. “We were very, very poor,” she later said. He died on Dec. 12, 1981, of what prosecutors said was an overdose of the anticonvulsant Dilantin. At the 2020 plea, his mother, Connie Weeks, told Jones she hoped she would live “a long and miserable life behind bars.”

The Case

The Death Shift

Jones was born about 1950 and grew up in northwest San Antonio, one of four children adopted by Richard and Gladys Jones. She finished vocational nursing school in 1977 and on Oct. 30, 1978, began work in the pediatric intensive care unit at the Bexar County Hospital, later renamed Medical Center Hospital (Texas Monthly; ProPublica).

By the fall of 1981 nurses had noticed that children kept crashing on her shift, the 3-to-11 p.m. shift. In October a nurse, Suzanna Maldonado, told the head nurse that “too many babies were dying in the ICU.” A later internal review counted 10 children who died between May and December 1981 after “sudden and unexplained” complications, with Jones present each time (Texas Monthly).

The hospital moved cautiously. In January 1982 a records review found no wrongdoing, and the hospital’s malpractice lawyer warned that firing Jones or calling the district attorney without proof was legally risky. An outside panel of six specialists, led by an anesthesiologist from Toronto’s Hospital for Sick Children, reviewed the unit in February 1982. Its report was sharply critical of how the unit was run and recommended that the head of the unit and the head nurse be relieved of their duties. It makes no mention of her name. In early March 1982 the hospital announced it would replace all of the unit’s vocational nurses with registered nurses. Jones quit on March 17, before the deadline, and the unexplained emergencies stopped (Texas Monthly). The hospital gave her a good recommendation and offered her another job (ProPublica).

A study by the federal Centers for Disease Control later documented the pattern without naming her. From April 1981 to June 1982, 34 of 42 deaths in the unit, 81 percent, occurred on the evening shift, against 36 of 106 over the previous four years. “[T]he presence of one nurse” was associated with a relative risk of death of 10.7 and of cardiopulmonary arrest of 25.5, and for unexpected clinical events the reported relative risk was infinite. The study drew no conclusion about the cause. “Although the cause of the epidemic remains unclear,” it said, surveillance of deaths and resuscitations “may allow early recognition of similar problems in other hospitals” (Istre et al., N. Engl. J. Med. 1985). On Jan. 13, 1983, a hospital committee, aware that a grand jury investigation was under way in Kerrville, recommended “judicious silence.” Dr. Vincent DiMaio, the county’s former medical examiner, said none of the deaths had been reported to his office: “Nobody reported any case to me from the hospital or the medical school.”

Kerrville

Dr. Kathleen Holland opened her pediatric practice in Kerrville on Aug. 23, 1982, and hired Jones. Kerr County charged her with Chelsea’s murder and seven counts of injury to a child involving other clinic patients; Texas Monthly counted seven patients and eight emergencies in 31 days (Jones v. State, 716 S.W.2d at 144-45; Texas Monthly).

Chelsea McClellan was one of them. On Aug. 24 her mother brought her in for what she thought was a cold, and after Jones took her into another room, Chelsea had what was described as a seizure and spent nine days in the hospital. On Sept. 17 Petti brought Chelsea’s brother in with the flu, and Dr. Holland decided Chelsea should get her routine shots. After two injections she went limp. Petti testified that the child’s arms “began flopping around” and her eyes “got all strange looking.” She stopped breathing again on the way to San Antonio and was pronounced dead at Comfort Community Hospital. Id.

A bottle of the paralytic drug Anectine, a brand of succinylcholine, had gone missing from the clinic. According to the appeals court, it was later found with unexplained holes in its stopper, and Holland turned it over to the hospital administrator. Id. Trial testimony, as UPI reported it, accounted for the bottles differently: of three ordered, one was never opened, one disappeared and was never accounted for, and the third had lost five-sixths of its contents and been refilled with a saltwater solution. A physician testified that the missing amount, one vial plus five-sixths of another, was enough to paralyze four average-size men (UPI). Texas Rangers were notified on Sept. 24, 1982, and three days later Holland turned over what she had found and told Jones not to come back (KENS 5).

The Investigation

Chelsea’s body was exhumed on May 7, 1983 (KENS 5). The appeals court’s own account recites that in May 1983 a physician gave the State’s toxicologist tissue samples he had maintained, and that the body was then exhumed and additional samples taken (Jones, 716 S.W.2d at 145). That spring a Kerr County grand jury indicted Jones for murder, along with seven counts of injury to a child involving other clinic patients (Texas Monthly). Jones denied everything. “I haven’t killed a damn soul,” she told Texas Monthly, and she blamed the doctors. In San Antonio, District Attorney Sam Millsap was investigating the hospital deaths and the hospital’s handling of them, and a Bexar County grand jury later indicted her in the Rolando Santos case.

The Georgetown Trial

Because of pretrial publicity, the trial was moved to Georgetown (UPI). Jury selection began on Jan. 16, 1984 (UPI). Kerr County District Attorney Ron Sutton called 44 witnesses over 14 days (UPI). Petti McClellan testified. So did Dr. Fredric Rieders, who told the jury, “I detected, identified and measured succinylcholine.” A former roommate who was also a vocational nurse testified that she and Jones had researched the drug (UPI). Jones did not testify (UPI; ProPublica).

On Feb. 15, 1984, after about four and a half hours of deliberation, the jury convicted her. The next day, after little more than an hour, it assessed 99 years. The jury was seven women and five men. “Essentially the facts were so overwhelming we could not ignore them,” said the foreman, Edwin Edwards of Georgetown. Chelsea’s grandmother, Robin Alexander, said: “We can finally bury her and they can dig her up no more.” On Oct. 24, 1984, state District Judge Pat Priest found her guilty of injury to a child in the Santos case, in a bench trial, and sentenced her to 60 years, to run concurrently.

The Prison Years

Jones was denied parole again and again (San Antonio Report). In 2013, Houston television station KPRC reported that she was nonetheless due to walk free in 2018 under the mandatory supervision law in effect when she was sentenced. “Texas is going to release a serial killer of children,” Petti McClellan said (KPRC). The families organized, and Petti told reporters, “This is my mission now.” She added: “Losing a child does not consume you; it drives you.” Bexar County District Attorney Susan Reed called the release “outrageous” and reopened the files.

In a March 2011 letter to the Texas Board of Nursing, Jones wrote: “I look back now on what I did and agree with you that it was heinous, that I was heinous.”

The New Prosecution

Assistant District Attorney Jason Goss took the case in December 2016 and had 20 boxes of old files moved into an unmarked office. By the end the file had grown to 42 boxes. “The documents are old. There’s no smoking gun. There’s no confession,” he said (ProPublica). Many of the hospital’s records were gone. “If we had the medical records that would be a different proposition,” he said (Texas Public Radio).

The records were not merely lost. Texas Public Radio reported that the hospital shredded medical records while they were under a court order to preserve them, and the district attorney’s office confirmed that the shredding order was signed the same day the hospital’s executive director resigned. That is why the 1981 and 1982 deaths could not be charged on the medical proof, and it is the clearest line in this case between an institution protecting itself and a prosecution that arrived 35 years late.

Joshua Sawyer’s mother had kept his medical records for 30 years. They showed a Dilantin level of 59.6 micrograms per milliliter, against a therapeutic range of 10 to 20 (ProPublica). Laboratories treat 30 as a critical value, and the World Health Organization’s poisons monograph lists 50 to 70 as a band in which fatalities have been recorded. The chief medical examiner amended Joshua’s death certificate to homicide, and listed four other children’s deaths as undetermined (Texas Monthly). Between May and October 2017, Bexar County grand juries returned five murder indictments (ProPublica). Texas Monthly reported that the grand jury deliberated on Joshua’s case for less than two minutes.

The defense argued that the decades of delay violated Jones’s rights to due process and a speedy trial. Judge Frank Castro rejected the speedy trial claim at an April 2018 hearing (ProPublica), and the court of appeals later affirmed the denial of her motion to dismiss on both grounds (Jones v. State, No. 04-20-00108-CR). At a pretrial hearing in 2018, Goss related that a retired parole officer recalled Jones telling her in 1998, “I really did kill those babies.” The officer had not recorded the words; her note at the time said only that Jones “admits she committed the present offenses” (Texas Monthly).

On Jan. 16, 2020, Jones pleaded guilty to murdering Joshua Sawyer with a deadly weapon and was sentenced to life. The sentence requires her to serve 20 years before she is first eligible for parole, and with credit for the time she had already spent in the Bexar County jail that is about 18 years from the plea. She confirmed the plea “in a raspy whisper” and otherwise said nothing (Texas Monthly). The four other cases were dismissed. The sentence “doesn’t come close to what you did,” Judge Castro told her (Texas Monthly).

How Many?

Jones has been convicted in two deaths. The numbers beyond that are estimates, and they vary widely. The CDC study counted 42 deaths in the unit during the period it examined, and ProPublica reported that Jones had cared for 20 of those children (ProPublica). Peter Elkind, who has written about the case since 1983, says she is credibly suspected in “more than a dozen” deaths and routinely blamed for “up to sixty” (Texas Monthly). Prosecutors have put the number at about 20 (NBC News), 47 to 60 (Texas Public Radio) and as many as 60.

How the Case Was Built

Succinylcholine is hard to find after death. In living subjects it is usually detectable in blood for only about 10 minutes after an injection. The body breaks it down rapidly, by plasma cholinesterase, into succinylmonocholine, then more slowly into succinic acid and choline (FDA label). Succinic acid and choline are both normal constituents of biological tissue, so neither one is evidence of an injection, which is why modern casework targets succinylmonocholine instead. The same work cuts the other way too: its authors found succinylmonocholine at 10 to 30 nanograms per gram in postmortem liver and kidney from people who had never received the drug (Rieders et al., 2004). The McClellan case needed a way to find the drug itself.

• The exhumation. Chelsea’s body was exhumed in May 1983, and tissue samples went to toxicologist Dr. Fredric Rieders.
• The method. Rieders tested the tissue at the Karolinska Institute in Stockholm, using an ion-pair extraction and gas chromatography-mass spectrometry method that the appeals court found Dr. Bo Holmstedt had developed. The method pulled the drug out of tissue as an ion pair, converted it chemically to a tertiary amine, and identified and measured that amine against a deuterated internal standard, down to 5 nanograms per gram. It looked for intact succinylcholine, not its breakdown products. The method had been published in 1983, with Holmstedt the fourth of six authors.
• The confirmation. Holmstedt re-ran the tissue himself and reached the same result, and a third toxicologist reviewed the data and agreed.
• The defense attack. The defense called Holmstedt himself, a proponent of the test, on direct examination. He acknowledged that the method had been developed on embalmed rat tissue, that the human-tissue work was unpublished because the case in which it was used was still pending, and that he knew of no one but a collaborator, Dr. Forney, who had duplicated it. Defense toxicologist Peter Proctor testified that independent reproduction was essential. The defense also suggested the body could produce the result on its own, which Rieders called highly improbable.
• The answer to the rat tissue. The court’s recitation of the rat-tissue point came attached to the rest of the sentence: the test developed on embalmed rat tissue “has been published in two major toxicology journals.” One of those publications is why the result mattered here. In 1982 the same group reported that intact succinylcholine could still be identified and measured in the embalmed kidney, liver and muscle of injected rats after six months of storage, with the highest concentrations at the injection site. That is published, pre-trial support for the one thing this case required, which is that the drug survives in embalmed tissue.

Later research shows why the defense questions were fair ones. Studies have found succinylmonocholine in normal postmortem tissue and in decomposed bodies that never received the drug, and have found an unidentified substance in ordinary serum and urine that co-elutes with it and mimics its main ion transition (Ballard 2006; Kuepper 2008; Kuepper 2011). Each of those is a reason to distrust a result for the breakdown product rather than for the intact drug the McClellan test targeted, so none of them undoes the 1984 finding. The honest complication sits in the same 2011 paper. In a known fatal succinylcholine poisoning the authors found neither the drug nor its breakdown product in any tissue, and concluded that tissue does not permit a reliable diagnosis because both compounds are unstable in tissue that contains esterases. That is a reason a modern laboratory might find nothing. It is not a reason a 1984 laboratory would find something that was not there, and the embalmed-tissue work is what changes the calculation for an exhumed body. All of it shows how much a postmortem drug finding depends on what exactly was measured.

The Signature Question

What tied Jones to the children was not a distinctive method but a pattern of opportunity: emergencies clustered on her shifts and at her new job. In the Santos appeal, the San Antonio court upheld the admission of testimony from Dr. Gregory Istre, the epidemiologist who had led the federal study, that a child’s risk of dying was 10.7 times greater while Jones was working. Two things limit that holding. The Santos case was tried to the judge rather than to a jury, and the court leaned on it: “The trier of fact in this case was the experienced trial judge himself.” And the chief justice, concurring, said evidence of those prior events had “no probative value whatever,” resting on the same posture, that a reviewing court must presume the judge did not consider inadmissible testimony. Jones v. State, 751 S.W.2d 682, 684, 686, 688 (Tex. App.-San Antonio 1988).

A pattern like that is powerful, but it is not a signature in the Rule 404(b) sense. Under Segundo v. State, 270 S.W.3d 79, 88-90 (Tex. Crim. App. 2008), identity evidence must be distinctive enough to mark crimes as one person’s work. Being present when things go wrong shows opportunity. In the McClellan case, the chemistry carried identity: the test put a specific drug in a specific child after a specific injection.

Would the 1984 test pass Kelly today? Every gap the defense raised maps onto a Kelly factor: acceptance in the field, published literature, and the availability of other experts to test and evaluate the technique. A court today would weigh them rather than stopping at general acceptance. Whatever the answer, the State would have to persuade the court by clear and convincing evidence that the evidence was reliable and therefore relevant.

What the Defense Could Have Tested

• Independent replication. Ask for the analysis to be repeated by a laboratory with no stake in the method.
• Controls. Ask for results on tissue from people who never received the drug, handled the same way as the exhumed samples.
• The full data. Ask for the instrument data, chromatograms and spectra behind each result, not only the conclusion.
• What was measured. Pin down whether the result reflects the intact drug or a breakdown product, and what else could produce the same signal.
• The pattern evidence. Test the statistics behind any shift analysis, including how cases were selected and what the baseline rate was.

How It Works in Texas

Reliability. Kelly governs all scientific evidence offered under Rule 702. Hartman v. State, 946 S.W.2d 60, 63 (Tex. Crim. App. 1997). The defense raises the challenge by objecting and asking for a hearing outside the jury’s presence. See Challenging the Science: The Kelly Reliability Hearing and How to Challenge a Forensic Expert Witness in Texas. Novel science keeps arriving in Texas courtrooms; see A Virus as a Witness and Bite Marks, Hair Analysis, and the Science Courts Are Still Admitting.

Accreditation. Under Art. 38.35 of the Code of Criminal Procedure, a forensic analysis means a “medical, chemical, toxicologic, ballistic, or other expert examination or test performed on physical evidence,” and that analysis and the expert testimony relating to it are inadmissible if, at the time of the analysis, the laboratory was not accredited by the Forensic Science Commission. The one exception saves a laboratory that was eligible for accreditation when the test was run and obtains accreditation before testimony. Several things sit outside the definition altogether, including latent print examination, breath tests and digital evidence, and the article does not reach the portion of an autopsy performed by a licensed physician. Accreditation is a floor, not a guarantee; see Why Lab Accreditation Does Not Guarantee Your Test Results Are Correct and Who Checks the Crime Labs?

Discovery. For offenses committed on or after Jan. 1, 2014, Art. 39.14 requires the State, as soon as practicable after a timely defense request, to produce and permit the inspection and copying of offense reports, designated documents and papers, witness statements and designated tangible things that constitute or contain evidence material to any matter involved in the action. It is a duty rather than a courtesy, and it reaches items held by the State or by any person under contract with the State, which is the hook for an outside laboratory. Older cases fall under the former, narrower version of the article, plus Brady. For the fight over its scope, see The Michael Morton Act, the Richard Miles Act and the Heath Opinion Now Under Fire.

After conviction. Art. 11.073 applies to relevant scientific evidence that either was not available to be offered at trial or contradicts scientific evidence the State relied on at trial. Relief requires specific facts showing that the evidence is available now and was not ascertainable through reasonable diligence before or during trial, that it would be admissible under the rules of evidence at a trial held on the date of the application, and that had it been presented at trial, on the preponderance of the evidence the person would not have been convicted. On the diligence question the court must consider whether the field of scientific knowledge, a testifying expert’s scientific knowledge, or the scientific method itself has changed since trial. Note the words. A testifying expert’s scientific knowledge is not the same as an expert’s opinion, and that distinction is what the State argues. See Texas’s Junk Science Writ.

The case still turns up in Texas courts. In 2014 the San Antonio court cited Jones while reviewing expert testimony in the case of a dialysis nurse accused of injecting bleach into patients’ lines, citing it for the use of gas chromatography-mass spectrometry to extrapolate from a rat-tissue study to human-tissue testing. That opinion was later vacated on a jury-charge question and the court affirmed again on remand, so it is worth citing for its reasoning rather than as binding authority. Saenz v. State, 421 S.W.3d 725, 762 (Tex. App.-San Antonio 2014), vacated on other grounds, 451 S.W.3d 388 (Tex. Crim. App. 2014), aff’d on remand, No. 04-12-00238-CR (Tex. App.-San Antonio Aug. 26, 2015).

Where Things Stand Now

The release law that nearly freed her is what the prison system calls mandatory supervision, release “when the combination of actual calendar time and good conduct time equal the sentence.” For offenses on or after Sept. 1, 1996 the parole board may deny it case by case. For hers it could not (TDCJ). Her appeal, which argued the decades of delay, was rejected in 2021. Jones v. State, No. 04-20-00108-CR (Tex. App.-San Antonio Aug. 25, 2021) (mem. op., not designated for publication).

As of October 2026, TDCJ lists her at the Christina Melton Crain Unit with a parole eligibility date of Dec. 3, 2037. Elkind’s book on the case, The Death Shift, was first published in 1989, and San Antonio’s KENS 5 revisited the case in its Vile podcast.

Sources

This post describes Genene Jones’s convictions using court records and published reporting. Deaths she has not been convicted of causing are described as estimates or claims, not facts, and the other 2017 indictments were dismissed. This post is offered for general information about Texas law and forensic evidence, is not legal advice, and does not create an attorney-client relationship.