A number of lethal injection executions have gone terribly, visibly wrong. Michael Radelet, a professor of sociology and law, has compiled a list of thirty-six botched executions, which he defines as executions where there is the appearance of prolonged suffering on the part of the condemned inmate for twenty minutes or more.217 Because states do not make public, maintain, or even keep records of their executions (see the U.S. Constitutional Law section of Chapter Seven), this list was developed from media reports. There may be other botched executions that were never reported. In addition, there is no way to know how many prisoners killed by lethal injections suffered needlessly, but invisibly, because of inadequate anesthesia masked by a neuromuscular blocking agent.
Lethal injection executions where the condemned inmates suffering was visible to the witnesses include:
The medical team could not find a suitable vein. Now I was really beginning to worry. If you cant stick a vein then a cut-down [where a cut is made into the vein to insert the chemicals] has to be performed. I have never seen one and would just as soon go through the rest of my career the same way. Just when I was really getting worried, one of the medical people hit the vein in the left leg.228
Because of recent litigation in North Carolina challenging that states lethal injection protocol, evidence of a number of botched executions in that state have recently become public:
Even when lethal injections have appeared to proceed smoothly, however, they may nonetheless have involved considerable pain and suffering. The inability to ascertain whether or not more prisoners have suffered during their executions stems from the use of pancuronium bromide, which prevents the prisoners from communicating verbally or physically what they are experiencing. Witnesses to the execution see a person lying quietly; they have no way of knowing whether he is in fact properly anesthetized or whether he is experiencing excruciating pain behind his paralyzed face.
Execution recordse.g., execution logs, autopsies, and toxicology reportsare necessary to conduct accurate post-mortem reviews of how the execution proceeded, including whether the prisoner reached an appropriate level of anesthesia.233 But corrections agencies have refused to create or keep such records, and agencies have refused to make them publicly available when they have been created or kept.234 For example, Texas, which has conducted 362 lethal injection executions, the most in the United States,235 stopped conducting autopsies of its executed prisoners in 1989.236
Execution logs from Californiathe only state in which such records have been made publicly available, and only because of litigationstrongly suggest that lethal injection executions in that state are not going according to plan. When a barbiturate like sodium thiopental is used during surgery, the patient goes limp within seconds after the drug begins flowing into his veins.237 He may take a few breaths, cough, hiccup, or have some erratic breathing, but there would be no regular and ongoing up and down chest movements.238 The anesthesia removes the patients ability to breathe on his own, which is why a doctor will intubate him so that a machine can do his breathing for him during surgery.239 Yet in California, six recent lethal injection execution logs indicate that prisoners were breathing more than a minute after they should have received a dose of sodium thiopental ten times that used in surgery.240 According to the execution logs:241
The logs do not prove that these six men were conscious when the pancuronium bromide and potassium chloride were injected. But the fact that their breathing did not stop when expected suggests adequate doses of sodium thiopental may not have been administered. At the very least, the logs point to the importance in three-drug lethal injection executions of having someone present to establish the level of anesthesia before the second and third drugs are administered.
Eyewitness testimony about lethal injection executions in Texas also raises concerns some prisoners in Texas were breathing after the administration of the sodium thiopental should have paralyzed their lung muscles. Reverend Carroll Pickett witnessed ninety-five lethal injection executions in Texas from 1982 through 1995.244 As the condemned inmates spiritual advisor on the day of his execution, Pickett stayed with the inmate throughout the execution until the inmate died. Once the inmate was on the gurney, Pickett stood next to him, his right hand touching the inmates right knee. During some of the executions, he saw some of the boys with their eyes open and looking at me after the thiopental came, I felt like I let [the prisoner] down, because the execution was not proceeding exactly as I told [the prisoner].245 Human Rights Watch asked Pickett if he signaled anything to the warden when he noticed a prisoners eyes open. He said no, that it had not been clear to him that something was wrong.246 When asked if he remembered any of the inmates breathing after the administration of the sodium thiopental, Pickett said that he did not see any of them stop breathing after that. That just put them to sleep. But they kept breathing. All of them.247
Pickett did not have any medical training; he had not been asked to monitor the condemned inmates breathing; and the executions were many years ago. Nevertheless, his memory of open eyes and breathing prisoners suggests there in fact may have been serious problems with the way Texas executed its prisoners.
 See Michael Radelet, Post-Furman Botched Executions, http://deathpenaltyinfo.org/article.php?scid=8&did=478 (retrieved April 5, 2006). Also: Human Rights Watch telephone interview with Michael Radelet, professor of sociology, University of Colorado at Boulder, Boulder, Colorado, March 1, 2006.
 Murderer of Three Women is Executed in Texas, p. 9.
 Landry Executed for 82 Robbery-Slaying, Dallas Morning News, December 13, 1988, p. 29A.
 Witness to an Execution, Houston Chronicle, May 27, 1989, p. 11.
 Niles Group Questions Execution Procedure, United Press International, November 8, 1992.
 Joe Farmer, Rector, 40, Executed for Officers Slaying, Arkansas Democrat-Gazette, January 25, 1992, p. 1B; Sonja Clinesmith, Moans Pierced Silence During Wait, Arkansas Democrat Gazette, January 26, 1992, p. 1B; Marshall Frady, Death in Arkansas, The New Yorker, February 22, 1993, p. 105.
 Rob Karwath and Susan Kuczka, Gacy Execution Delay Blamed on Clogged IV Tube, Chicago Tribune, May 11, 1994, p. 1 (Metro Lake Section).
 Tim ONeil, Too-Tight Strap Hampered Execution, St. Louis Post-Dispatch, May 8, 1995, p. 6B.
 Sheri Edwards and Suzanne McBride, Doctors Aid in Injection Violated Ethics Rule: Physician Helped Insert the Lethal Tube in a Breach of AMAs Policy Forbidding Active Role in Execution, Indianapolis Star, July 19, 1996, p. A1.
 Killer Helps Officials Find A Vein At His Execution, Chattanooga Free Press, June 13, 1997, p. A7.
 1st Try Fails to Execute Texas Death Row Inmate, Orlando Sentinel, April 23, 1998, p. A16; Michael Graczyk, Texas Executes Man Who Killed San Antonio Attorney at Age 17, Austin American-Statesman, April 23, 1998, p. B5.
 Sarah Rimber, Working Death Row, New York Times, December 17, 2000, p. 1.
 Rhonda Cook, "Gang leader executed by injection Death comes 25 years after boy, 11, slain" Atlanta Journal Constitution, November 7, 2001, p. B1.
 Order. Brown v. Beck, No. 5:06-CT-3018-H, April 7, 2006, p. 9.
 Ibid., p. 10.
 In April of 2005, a team of medical doctors reported in the British medical journal The Lancet that toxicology reports on forty-three of forty-nine executed inmates revealed the anesthetic administered during lethal injections was lower than that required for surgery. Indeed, in twenty-one of the inmates, the concentrations of thiopental in the blood were consistent with awareness. The report concludes, "Failures in protocol design, implementation, monitoring and review might have led to the unnecessary suffering of at least some of those executed. Because participation of doctors in protocol design or execution is ethically prohibited, adequate anesthesia cannot be certain. Therefore, to prevent unnecessary cruelty and suffering, cessation and public review of lethal injection is warranted." G. K. Leonidas, et al., Inadequate Anesthesia in Lethal Injection for Execution, The Lancet, Vol.365 (9468), April 16, 2005, p.1412. Medical experts have subsequently discredited the Lancet report because the blood used in the toxicology analysis was drawn many hours after the execution. To be most accurate, blood used for a toxicology analysis would have to be drawn soon after the prisoners death. See, e.g., Study: Lethal Injection Not Painless, Chicago Tribune, April 15, 2005, http://www.med.miami.edu/communications/som_news/index.asp?id=470 (retrieved April 2, 2006).
 See, for instance, policies in Missouri, Louisiana, and North Carolina: Missouri does not keep any records from its executions (Defendant Crawfords Answers to Plaintiffs First Interrogatory, Taylor v. Crawford, Case No. 05-4173-CV-C-SOW, September 12, 2005, p. 24); Louisiana does not keep its execution records for more than five years (Inglis Deposition, p. 57); North Carolina does not keep any execution records either (Testimony of Polk, p.114).
 DPIC, Execution Database.
 Harris v. Johnson, et al., April 15, 2004, p. 5.
 Interview with Heath, March 6, 2006; Interview with Dershwitz, March 1, 2006.
 E-mail correspondence to Human Rights Watch from Dershwitz, March 9, 2006.
 Doses of sodium thiopental used in surgery are typically one-tenth the five grams called for in Californias lethal injection executions at the time of these six executions. San Quentin Procedure No. 770, p. 32. California has since changed its dosage of sodium thiopental, from five grams to 1.5 grams. See Chapter Five on the Morales v. Hickman case.
 It is unclear who was responsible for keeping the execution log and what the protocol was for determining when respirations ceased. E-mail correspondence to Human Rights Watch from John Grele, attorney for Morales, April 1, 2006. (Copies of the six execution logs are on file with Human Rights Watch.)
 The execution log states that Richs respirations ceased at 12:08 a.m., but notes that Rich had chest movements lasting from 12:09 to 12:10 a.m. These chest movements began after Rich had supposedly stopped breathing and three minutes after the administration of the thiopental. The chest movements are consistent with an attempt to fight off the accruing paralytic effect of the pancuronium. Third Declaration of Dr. Mark Heath, Morales v. Hickman, February 9, 2006, p. 6.
 The records are inconsistent. The formal execution log suggests that Williams stopped breathing at 12:28 a.m. and indicates that potassium chloride was injected at 12:32 a.m., whereas the execution teams log states that Williams stopped breathing at 12:34 a.m., when the potassium chloride was injected. It appears that the formal log was altered without any indication as to who made the alteration.
 Interview with Pickett.