Of primary importance in this case is the following question: under what circumstances should a surgeon in charge of an operation be held vicariously liable for the negligence of medical personnel who assist in performing the operation?
On 1 June 1981, Mrs. Etta Harris underwent back surgery under general anesthesia at Beaufort County Hospital. Defendant George Miller, M.D., an orthopedic surgeon, performed the surgery assisted by William Hawkes, a nurse anesthetist. As a result of inadequate oxygenation during the surgery, Mrs. Harris suffered brain damage and paralysis. Some six years later, she died from complications secondary to the brain damage.
*383In 1983, Mrs. Harris and her husband filed suit for personal injury and loss of consortium against the hospital, Dr. Miller and Nurse Hawkes. They later settled their claims against the hospital and Nurse Hawkes, executing a covenant not to sue, but specifically reserved the right to pursue any claims against Dr. Miller. When Mrs. Harris died, the complaint was amended to allege wrongful death and the case proceeded against Dr. Miller on theories of direct and vicarious liability. Plaintiff claimed that Dr. Miller was negligent in causing a severe bleeding problem during the surgery, in failing to properly treat the bleeding problem and in failing to adequately supervise Nurse Hawkes. Plaintiff also claimed that Dr. Miller should be held liable for the negligence of Nurse Hawkes under the doctrine of respondeat superior. At the close of plaintiff’s evidence, the trial court granted Dr. Miller’s motion for a directed verdict on the vicarious liability claim, finding the evidence insufficient to establish a master-servant relationship between Dr. Miller and Nurse Hawkes. As an alternative basis for the directed verdict, the trial court held that the prior release of Nurse Hawkes served to exonerate Dr. Miller. The case was submitted to the jury on the sole issue of Dr. Miller’s negligence.
The jury found for Dr. Miller and plaintiff appealed. The Court of Appeals, Judge Phillips dissenting, affirmed the judgment below. 103 N.C. App. 312, 407 S.E.2d 556 (1991). Plaintiff appealed to this Court on the basis of the dissent, presenting the following issues: 1) whether the trial court erred in directing a verdict on the issue of Dr. Miller’s vicarious liability, and 2) whether the trial court erred in excluding certain testimony of plaintiff’s expert on nurse anesthesia care. We granted plaintiff’s petition for discretionary review of an additional issue, not addressed by the Court of Appeals: whether the trial court erred in ruling that the release of the servant, Nurse Hawkes, extinguished the vicarious liability of the master, Dr. Miller. Having concluded that the Court of Appeals erred in affirming the trial court’s rulings on the first two issues, and that the trial court erred in its ruling on the release issue, we now reverse the Court of Appeals and grant plaintiff a new trial.
Plaintiff’s evidence tended to show the following.
In early 1981, Mrs. Harris began experiencing severe back pain. She consulted Dr. Miller, an orthopedic surgeon, who diag*384nosed a ruptured disc requiring surgery. Dr. Miller performed the surgery on June 1, 1981, at Beaufort County Hospital, where he had staff privileges. Anesthesia was administered by Nurse Hawkes, a certified registered nurse anesthetist employed by the hospital and assigned to the case by the hospital’s Chief Anesthetist. Because the hospital did not employ a staff anesthesiologist, Nurse Hawkes worked for the duration of the case, as stated in the hospital’s Anesthesia Manual, under the “responsibility and supervision” of Dr. Miller. No anesthesiologist was available for consultation within thirty miles.
The operation appears to have been doomed from the start by Nurse Hawkes’ negligent performance of the pre-operative anesthesia evaluation. Among other errors, Nurse Hawkes interpreted Mrs. Harris’ chest X-rays as “negative” when in fact she had an enlarged heart — evidence of past heart disease —and failed to perform an electrocardiogram despite her mild obesity and history of high blood pressure. As a result, Nurse Hawkes was unaware of Mrs. Harris’ heart problems, an unfortunate circumstance given that he would be using anesthetic agents — Demerol, Innovar and Ethrane — that can significantly lower blood pressure in patients with depressed cardiac function.
Come the day of the operation, Nurse Hawkes put Mrs. Harris to sleep at 7:45 a.m. After inserting an endotracheal tube, he turned the patient and started the maintenance anesthesia: sixty-six percent nitrous oxide, thirty-three percent oxygen and one percent ethrane. As expected, Mrs. Harris’ blood pressure dropped slightly. In most patients, the drop in blood pressure at induction is a normal reaction to the anesthetic agents and is no cause for concern; the blood pressure soon rights itself in response to the stimulation of surgery. However, when surgery began at 8:05 a.m., Mrs. Harris’ blood pressure did not return to normal. Instead, it continued to drop, while her pulse rate rose dramatically.
Thinking that his patient was feeling pain, too lightly anesthetized, Nurse Hawkes administered high dosages of Demerol and Innovar, and continued to give high levels of Ethrane. Her pulse rate did not decrease, however, and her blood pressure remained dangerously low. In actuality, Mrs. Harris was suffering from a lack of oxygen and too much anesthesia. Post-surgery X-rays revealed that the endotracheal tube had slipped into her right lung, leaving the left lung unventilated. Her heart was beating faster *385to compensate for the lack of oxygen, her blood pressure unresponsive because of the anesthetics. Nurse Hawkes did not realize that the endotracheal tube had slipped because, contrary to standard procedure, he had not checked for bilateral breath sounds when he turned the patient after intubation.
Nurse Hawkes continued to give high levels of anesthesia from 8 to 9 a.m. During this time, Mrs. Harris’ blood pressure remained at 100 systolic, 70 diastolic, some thirty to fifty points lower than normal, and her pulse rate at 130. Nurse Hawkes did not inform Dr. Miller of the problem.
For Dr. Miller, the operation proceeded smoothly until 8:40 a.m., when he noticed an unusual amount of bleeding. Having just finished removing the extruded disk, Dr. Miller applied small packs to the bleeding and proceeded to clean the disk space. This done, he removed the packing only to find that the bleeding had continued unabated. By 9 a.m., Mrs. Harris had lost roughly 400 cc’s of blood, 300 cc’s more than a patient would normally lose over the entire operation. At this point, Dr. Miller instructed Nurse Hawkes to start giving the patient blood.
In derogation of this direct order, Nurse Hawkes did not start giving blood until roughly 9:40 a.m. In the meantime, Mrs. Harris suffered a precipitous drop in blood pressure due to the loss of blood volume. By 9:15 a.m., her blood pressure had dropped to 90 systolic, 60 diastolic; by 9:25 a.m., to 80 systolic, her diastolic now inaudible; by 9:40 a.m., to 70 systolic. Here her blood pressure would remain, a level incompatible with normal brain function, until 10:20 a.m., while her pulse rate rose to 140. Yet still, Nurse Hawkes did not inform Dr. Miller. Nor did he take appropriate remedial measures. The proper course of action would have been to cut off all anesthesia and give one hundred percent oxygen as of 9:15, at the very latest. Instead, Nurse Hawkes administered a final dose of Demerol at 9:15 a.m., continued Ethrane until 9:45 a.m. and left the oxygen at thirty-three percent until 11 a.m., long after Mrs. Harris had suffered permanent brain damage.
All the while, Mrs. Harris continued to bleed. By 10 a.m., Dr. Miller had succeeded only in identifying the source of the bleeding, a small hole in one of the vertebra on which he had operated. Realizing, then, that there was a “major bleeding problem,” Dr. Miller requested more blood and applied Surgicel to the wound. He also called in another surgeon to assist him. The Surgicel *386stopped the bleeding. Unfortunately, Dr. Miller removed the Surgicel after only twenty minutes, fearing that it would swell and damage Mrs. Harris’ spinal cord. Though the bleeding resumed at a vigorous rate, Dr. Miller did not replace the Surgicel. Instead, he tried again to control the bleeding with direct pressure. According to one of plaintiff’s experts, in failing to keep the Surgicel in place, Dr. Miller turned “a very bad situation into a[n] irretrievable one.”
By 10:20 a.m., Mrs. Harris’ systolic blood pressure had dropped to 40, her diastolic blood pressure still inaudible. At this point, Nurse Hawkes administered neo-synephrine, a vasoconstrictor. Mrs. Harris’ blood pressure rose briefly but soon plunged again. Now her pulse rate began to drop as well, rapidly. When Nurse Hawkes checked her vital signs at 11:10 a.m., she had no discernible blood pressure or pulse. He then, for the first time, informed Dr. Miller that Mrs. Harris was in extremis. This intelligence came as a surprise to Dr. Miller. Though his patient had been losing blood at an alarming rate for two and a half hours, Dr. Miller had not once inquired about her vital signs.
Dr. Miller immediately made a partial closure of Mrs. Harris’ back, turned her and devoted himself to the resuscitation effort. With the help of a number of surgical specialists he had called in to assist him, Dr. Miller ultimately succeeded in restoring Mrs. Harris’ blood pressure and pulse. His efforts, however, came far too late to prevent her injuries, the damage to her brain having occurred sometime between 9 and 10:45 a.m.
Mrs. Harris remained in a coma for some time after the operation. When she regained consciousness, she could no longer move any of her limbs effectively. She could only breathe with the aid of a tube inserted in her neck, and paralysis of the vocal cords prevented her from speaking. She would spend eight months in rehabilitative hospitals before returning home. The home remodeled to meet her many needs, she was cared for almost exclusively by her husband for the next five years. Despite daily, agonizing, rehabilitation exercises, her health slowly deteriorated and she died on November 8, 1987.
The first issue raised is whether the trial court properly directed a verdict for Dr. Miller on plaintiff’s vicarious liability claim. We *387hold that plaintiff presented sufficient evidence of a master-servant relationship between Dr. Miller and Nurse Hawkes to get this claim to the jury.
 Because Nurse Hawkes was employed by Beaufort County Hospital, Dr. Miller’s vicarious liability, if any, depends upon the “borrowed servant” doctrine: One who borrows another’s employee may be considered a temporary master liable in respondeat superior for the borrowed employee’s negligent acts if acquiring the same right of control over the employee as originally possessed by the lending employer. Weaver v. Bennett, 259 N.C. 16, 129 S.E.2d 610 (1963); Restatement (Second) of Agency § 227 comment a (1957). Though the rule is clear, our courts have not always consistently applied it in determining the liability of surgeons for the negligence of operating room personnel. Our first task, then, must be to clarify the proper application of the borrowed servant rule in this context.
The traditional test of liability under the borrowed servant rule is as follows:
Whether a servant furnished by one person to another becomes the employe (sic) of the person to whom he is loaned [depends on] whether he passes under the latter’s right of control with regard not only to the work to be done but also to the manner of performing it. ... A servant is the employe (sic) of the person who has the right of controlling the manner of his performance of the work, irrespective of whether he actually exercises that control or not.
(Emphasis in original). Weaver, 259 N.C. at 28, 129 S.E.2d at 618 (quoting Mature v. Angelo, 373 Pa. 593, 97 A.2d 59 (1953)); see also Hodge v. McGuire and Fingleton v. McGuire, 235 N.C. 132, 69 S.E.2d 227 (1952); and Leonard v. Transfer Co., 218 N.C. 667, 12 S.E.2d 729 (1940).
Employment, of course, is a matter of contract. Thus, where the parties have made an explicit agreement regarding the right of control, this agreement will be dispositive. Producers Chemical Co. v. McKay, 366 S.W.2d 220, 226 (Tex. 1963). Absent such an agreement, inferences must be drawn from the circumstances surrounding the employment. Id. Facts considered relevant include whether the lent servant is a specialist, which employer supplies *388the instrumentalities used to perform the work, the nature of those instrumentalities, the length of the employment, the course of dealing between the parties, whether the temporary employer has the skill or knowledge to control the manner in which the work is performed and whether the temporary employer in fact exercises such control. Of these, the actual exercise of control is the most weighty. See Lewis v. Barnhill, 267 N.C. 457, 148 S.E.2d 536 (1966); Leggette v. McCotter, 265 N.C. 617, 144 S.E.2d 849 (1965); Weaver, 259 N.C. 16, 129 S.E.2d 610; Hodge, 235 N.C. 132, 69 S.E.2d 227; Beatty v. H.B. Owsley & Sons, Inc., 53 N.C. App. 178, 280 S.E.2d 484, disc. rev. denied, 304 N.C. 192, 285 S.E.2d 95 (1981); see also Restatement (Second) of Agency § 227 comment c. Absent evidence to the contrary, the original employer is presumed to retain the right of control. Restatement (Second) of Agency § 227 comment b; 53 Am. Jur. 2d Master and Servant § 415, at 428 (1970).
North Carolina appellate courts have only twice had the opportunity to test the liability of a surgeon for the negligence of operating room personnel under the borrowed servant rule: in Jackson v. Joyner, 236 N.C. 259, 72 S.E.2d 589 (1952), and later in Starnes v. Hospital Authority, 28 N.C. App. 418, 221 S.E.2d 733 (1976). In Jackson, plaintiff’s intestate, an eight-year-old girl, died after a tonsillectomy performed by the defendant Dr. Joyner as a result of anesthesia complications. The girl had a cold at the time of the operation and developed pneumonia during the surgery, the anesthetic — ether—apparently having caused the infection to spread to the lungs. Anesthesia was administered by a Nurse Hanson, who was employed by the hospital. Jackson v. Sanitarium, 234 N.C. 222, 223-24, 67 S.E.2d 57, 59-60 (1951), reh’g denied, 235 N.C. 758, 69 S.E.2d 29 (1952). The trial court instructed the jury that the defendant could not be held liable for the negligence of Nurse Hanson. This Court granted plaintiff a new trial on the vicarious liability claim, reasoning as follows:
The record discloses that the child’s mother in arranging for the operation contacted and engaged only Dr. Joyner. He in turn, after demurring to the mother’s suggestion that her family physician be engaged to give the anesthetic, arranged for the assistance of the nurses, including nurse Hanson, who administered the anaesthetic.
On this record the evidence is sufficient to justify the inference that during the time the child was being prepared *389for the operation and while the operation was in progress, Dr. Joyner, as surgeon in charge, had full power of control over the nurses, including nurse Hanson, so as to make him responsible for the way and manner in which the anaesthetic was administered by Hanson.
There is no discussion in Jackson of the facts underlying the Court’s conclusion that Dr. Joyner had the right to control Nurse Hanson other than the facts that Dr. Joyner, as opposed to the hospital, was engaged to perform the operation, and that he himself “arranged” for the assistance of the nurses. These facts do not sufficiently elucidate the right of control issue. The Court appears to have presumed that Dr. Joyner enjoyed the right of control from the mere fact that he was the “surgeon in charge.” Such a presumption runs contrary to the borrowed servant doctrine, part of which is that the lender rather than the borrower is presumed to have the right of control.
Though the presumption that the surgeon in charge controls all operating room personnel may have been appropriate in an era in which hospitals undertook only to “furnish room, food, facilities for operation, and attendance,” not to treat patients, Smith v. Duke University, 219 N.C. 628, 634, 14 S.E.2d 643, 647 (1941), and in which only physicians had the expertise to make treatment decisions, Byrd v. Hospital, 202 N.C. 337, 341-42, 162 S.E. 738, 740 (1932), it is no longer appropriate in this era. First of all, hospitals are now in the business of treatment. As stated in Rabon v. Hospital:
“The conception that the hospital does not undertake to treat the patient, does not undertake to act through its doctors and nurses, but undertakes instead simply to procure them to act upon their own responsibility, no longer reflects the fact. Present day hospitals, as their manner of operation plainly demonstrates, do far more than furnish facilities for treatment. They regularly employ on a salary basis a large staff of physicians, nurses and internes, as well as administrative and manual workers, and they charge patients for medical care and treatment, collecting for such services, if necessary, by legal action. Certainly, the person who avails himself of ‘hospital facilities’ expects that the hospital will attempt to cure him, not that its nurses or other employees will act on their own responsibility.”
Accordingly, hospitals now exercise significant control over the manner in which their employees, including staff physicians, provide treatment. This is done through hiring criteria, training, formal practice guidelines,1 hierarchical supervision structures, peer review groups and disciplinary measures. Stewart R. Reuter, Toward a More Realistic and Consistent Use of Respondeat Superior in the Hospital, 29 St. Louis U. L.J. 601, 632-34 (1985) [hereinafter Reuter]. As stated in Truhitte v. French Hospital, 128 Cal. App. 3d 332, 348-49, 180 Cal. Rptr. 152, 160 (1982): “Today’s hospitals hire, fire, train, and provide day-to-day supervision of their nurse-employees. . . . [H]ospitals can and do implement standards and regulations governing good surgery practices and techniques and are in the best position to enforce compliance.”
The degree to which hospitals may exercise control over the performance of operating room personnel is well illustrated in Sparger v. Worley Hospital, Inc., 547 S.W.2d 582 (Tex. 1977). At issue in that case was whether a surgeon should be liable for the negligent failure of operating room nurses to remove a surgical sponge from the plaintiff’s abdomen. In analyzing the right of control issue, the court noted that the hospital’s policy manual made accounting for sponges the responsibility of the nurses and mandated the following procedure. Nurses were required to count all sponges before the case began, recording the count in writing in the operative record, and again before closure of the operative incision. If the counts matched, a written note was to be made in the operative record and signed by the circulating nurse. If not, the surgeon was to be notified immediately and a search made. If the missing sponge was not found after a thorough search, an X-ray was to be taken. The X-ray could not be refused by the surgeon. Id. at 585, n.l. Thus, hospitals may retain absolute, authoritative control over the manner in which surgical procedures are performed.
While hospitals now exercise significant control over operating room personnel, surgeons are no longer the only experts in the operating room. The operating team now includes nurses, techni*391dans, interns, residents, anesthetists, anesthesiologists and other spedalized physicians. All of these are experts in their own fields, having received extensive training both in school and at the hospital. When directed to perform their duties, they do so without further instruction from the surgeon, relying instead on their own expertise regarding the manner in which those duties are performed. William H. Payne and K. Mike Mayes, Vicarious Liability and the Operating Room Surgeon, 17 S. Tex. L.J. 367, 387-90 (1976) [hereinafter Payne & Mayes]. Some of them, like anesthesiologists and technicians, may have expertise not possessed by the surgeon. Reuter, 29 St. Louis U. L.J. at 635-36. Thus, the surgeon will in some cases be ill-equipped, if not incapable, of controlling the manner in which assisting personnel perform their duties.
Even where the surgeon does have the knowledge or skill to control assisting personnel, it may be impractical for him to do so given the necessity of focusing on the surgical procedure. Generally, he has “no time to watch the anesthesiologist (-tist), nurses, or other assistants, much less direct them in the performance of their duties.” Payne & Mayes, 17 S. Tex. L.J. at 389-90; accord, e.g., Parker v. St. Paul Fire & Marine Ins. Co., 335 So. 2d 725, 734 (La. Ct. App.), writ refused, 338 So. 2d 700 (La. 1976) (evidence shows surgeon could not divert attention from surgical procedure to supervise transfusion of blood). As stated in Grant v. Touro Infirmary, 254 La. 204, 220, 223 So. 2d 148, 154 (1969): “[Operations performed under modern techniques require team performance, and the nurses and other personnel assisting in the operating room are not at all times under the immediate supervision and control of the operating surgeon . . . .”
Thus, today, the surgeon in charge may well have authority to direct only the tasks to be performed, not the manner of their performance.
 In light of the foregoing, we hold that surgeons should no longer be presumed to enjoy the authoritative control of a master over all who assist merely because they are “in charge” of the operation. Accord, e.g., Truhitte, 128 Cal. App. 3d 348, 180 Cal. Rptr. 152; Franklin v. Gupta, 81 Md. App. 345, 567 A.2d 524, cert. denied, 319 Md. 303, 572 A.2d 182 (1990); Thomas v. Hutchinson, 442 Pa. 118, 275 A.2d 23 (1971); Sparger, 547 S.W.2d 582; see also Payne & Mayes, 17 S. Tex. L.J. at 390. To the extent that Jackson sanctions such a presumption, we now overrule it.
*392The second North Carolina case to consider the liability of a surgeon for the negligence of operating room personnel was Starnes. In that case, a newborn was burned during a surgical procedure by an excessively hot water bottle which had been placed under him to keep him warm during the surgery. Warming the infant during surgery was the responsibility of the nurse anesthetist. Plaintiff alleged that the surgeon should be liable in respondeat superior for the negligence of the anesthetist. 28 N.C. App. at 419, 424, 221 S.E.2d at 735, 738. Affirming a directed verdict entered by the trial court in favor of the surgeon, the Court of Appeals explained that:
The department [of] anesthesiology assigned the anesthetist for the operation. Dr. Hamilton had no responsibility for her training or assignment. Absent some conduct or situation that should reasonably place the surgeon on notice of negligent procedure, we think the surgeon is entitled to rely on the expertise of the anesthetist. We find nothing to support general respondeat superior liability on the part of the surgeon.
Id. at 425, 221 S.E.2d at 738.
Given the facts presented, the court’s affirmance of the directed verdict is certainly reconcilable with traditional borrowed servant principles. That the hospital’s anesthesiology department trained its anesthetists indicates a retention by the hospital of the right to control those anesthetists. Nothing else appearing, it can only be inferred that the anesthetists remained the servants of the hospital while performing their surgical duties.
We do not approve, however, of the court’s rationale that a surgeon is entitled to rely on the expertise of the anesthetist absent reasonable notice of the anesthetist’s negligent conduct. If, by this language, the court meant that a surgeon, charged with the negligence of an assisting specialist, should only be held liable if he knew or should have known of the assistant’s negligent conduct, then the rationale is clearly inconsistent with the basic precept that respondeat superior imposes liability without fault. W. Page Keeton et al., Prosser and Keeton on the Law of Torts § 69 (5th ed. 1984). We assume, however, that the court understood this precept and intended instead to suggest a rule of limited liability for surgeons: At least where the negligence sought to be imputed is that of a specialist like an anesthetist, surgeons should be exempted altogether from respondeat superior liability and held respon*393sible only if they were negligent in supervising the specialist, on the assumption that surgeons never enjoy the right of control over assisting specialists.2 Though such a rule has. been adopted in some jurisdictions with regard to the surgeon’s liability for the negligence of anesthesiologists and anesthetists,3 we believe it unjustifiable.
While it may be true that surgeons generally rely on assisting specialists to perform their duties without supervision, and that assisting specialists may have the greater expertise in their particular fields, it is clearly not the case that surgeons never enjoy authoritative control over such assistants. As in the case at bar, the surgeon may have agreed with the hospital to control the performance of the specialist in question. Or a surgeon may know more about a particular procedure than an assisting specialist and actively supervise the latter’s performance, as' where the surgeon is assisted by a relatively inexperienced resident physician. Payne & Mayes, 17 S. Tex. L.J. at 389. Also, in emergency situations, the surgeon may have the right to control the performance of all life-saving acts: “Time is essential, and any loss of time could mean the life of the patient. The surgeon is implicitly given from the hospital, therefore, the right to control the nurses in regard to the ‘how’s’ of performing all acts in saving the patient.” Id. at 388, n.153.4 The testimony in the case at bar indicates that this right may also extend to the life-saving efforts of anesthetists and other specialists. And, finally, it is well-recognized that nurse anesthetists, when not supervised by an anesthesiologist, generally work under the direction and supervision of a physician. Margaret *394H. Meeker and Jane C. Rothrock, Alexander’s Care of the Patient in Surgery 147 (9th ed. 1991).
Nor is the fact that a borrowed employee has expertise considered a bar to respondeat superior liability in any other setting. See, e.g., Ward v. Gordon, 999 F.2d 1399 (9th Cir. 1993) (military physician held borrowed servant of private hospital where performed duties under supervision of hospital staff); Huff v. Marine Tank Testing Corp., 631 F.2d 1140 (4th Cir. 1980) (skilled welder held borrowed servant where lending employer equivalent of temporary employment agency and had no concern with details of welder’s work); United States v. N.A. Degerstrom, Inc., 408 F.2d 1130 (9th Cir. 1969) (experienced operator of heavy loading machine held borrowed servant where taught how to perform act causing damage by borrowing employer’s supervisor); Six Flags Over Georgia, Inc. v. Hill, 247 Ga. 375, 276 S.E.2d 572 (1981) (skilled ironworker held borrowed servant where, at least for particular work which caused injury, borrowing employer shown to be in complete control of how work was done); New York Central Railroad v. Northern Ind. Pub. Serv. Co., 140 Ind. App. 79, 221 N.E.2d 442 (1966) (crane operator held borrowed servant where foreman directing him an experienced crane operator himself); F.M. Pulliam v. Home Building Contractors, Inc., 363 S.W.2d 48 (Mo. Ct. App. 1962) (skilled carpenter held borrowed servant where borrowing employer paid his wages and had power to control how work should be done); Thompson v. Grumman Aerospace Corp., 78 N.Y.2d 553, 585 N.E.2d 355, 578 N.Y.S.2d 106 (1991) (experienced sheet metal mechanic held borrowed servant where recruited by lending employer specifically to work for borrower, where employee worked exclusively for borrower, and where lending employer had no knowledge or expertise regarding work performed by employee); Rorie v. Galveston, 471 S.W.2d 789 (Tex. 1971), cert. denied, 405 U.S. 988, 31 L. Ed. 2d 454 (1972) (experienced operator of complicated hoist held borrowed servant where borrower and lender expressly agreed employee would work under borrower’s control); and McKinney v. Air Venture Corp., 578 S.W.2d 849 (Tex. Civ. App. 1979) (experienced airplane pilot held borrowed servant where borrowing employer was owner of airplane and pilot’s first responsibility was safety of airplane). As is clear from the above-cited cases, the right of control may pass to a temporary employer, as a matter of fact, even where the borrowed employee has the skill of a specialist. When it does, respondeat superior liability must follow.
*395  Therefore, consistent with traditional agency principles, we hold that a surgeon may be held liable under the doctrine of respondeat superior for the negligence of even a skilled assistant if the surgeon in fact possessed the right to control that assistant at the time of the assistant’s negligent act regardless of whether the surgeon should reasonably have been aware of the negligent conduct sought to be imputed to him. Restatement (Second) of Agency § 227 comment a. To the extent that Starnes conflicts with this proposition, we now overrule it.
 In summary, we hold that a surgeon should not, as suggested by Jackson, be presumed to enjoy the authoritative control of a master merely because he is “in charge” of the operation. To the contrary, under traditional borrowed servant principles, the hospital must be presumed to retain the right of control over its operating room employees. Nor, however, should the surgeon be exempted from respondeat superior liability, as suggested by Starnes, merely because the negligence sought to be imputed is that of a skilled specialist. Whether a surgeon may be held vicariously liable for the negligence of one assisting in the operation depends on whether, in the particular case, the surgeon had the right to control the manner in which the assistant performed.
 Having outlined the proper reach of the borrowed servant doctrine in the context of the operating room, we turn now to the question of whether the trial court erred in granting Dr. Miller a directed verdict on plaintiff’s vicarious liability claim. To answer this question, we must decide whether plaintiff’s evidence of a temporary master-servant relationship between Dr. Miller and Nurse Hawkes was legally sufficient for his claim to be considered by the jury. United Laboratories, Inc. v. Kuykendall, 322 N.C. 643, 661, 370 S.E.2d 375, 387 (1988). Unlike the Court of Appeals, we believe it was.
Most persuasive is plaintiff’s evidence that Dr. Miller agreed with Beaufort County Hospital to control the performance of the nurse anesthetists, including Nurse Hawkes, assigned to his cases. See Producers Chemical Co. v. McKay, 366 S.W.2d 220, 226 (Tex. 1963) (“When a contract, written or oral, between two employers expressly provides that one or the other shall have right of control, solution of the [borrowed servant] question is relatively simple”). Such an agreement is indicated by the following language of the *396hospital’s Anesthesia Manual, whose provisions Dr. Miller agreed to comply with as a condition of his staff privileges:
Anesthesia care shall be provided by nurse anesthetists working under the responsibility and supervision of the Surgeon doing the case.
Administration of anesthesia shall be the sole responsibility of the Surgeon and anesthetist involved, and it shall be their responsibility to select and administer a proper agent with proper techniques.
Since the director [of the Anesthesia Department] is not an anesthesiologist, it is understood that the performance of anesthetists, while providing direct services to patients, shall be under the overall direction and supervision of the physician responsible for the patient’s care.
The Court of Appeals interpreted this language as vesting the surgeon with the right of supervision, not control: the power merely to point out the work to be done but not to direct the manner in which the work is to be performed. 103 N.C. App. at 324, 407 S.E.2d at 562-63. On the contrary, we believe this language, read as it must be in the light most favorable to plaintiff, Kuykendall, 322 N.C. at 661, 370 S.E.2d at 387, evinces a clear intent to transfer to the surgeon doing the case control over the manner in which the anesthetists performed their work. Indeed, the surgeon, under whose supervision the anesthetists are said to work, is given direct responsibility not only for the selection of anesthetic agents but also for the “techniques” used to administer them. Making decisions about agents and administration techniques — whether to preoxygenate the patient, whether to anesthetize using Demerol or a narcotic with fewer side effects, whether to maintain the patient strictly on inhalation anesthetics or on a combination of inhalation and intravenous anesthetics, whether to give the patient thirty-three or fifty percent oxygen — is the very essence of anesthesia practice. See 21 NCAC 36 .0226(c)(2)(B) (July 1993) (defining nurse anesthesia practice to include “selecting, implementing, and managing general anesthesia”). Clearly, then, the surgeon’s authority was over the manner of performance and not merely the tasks to be done. Compare Franklin, 81 Md. App. 345, 366, 567 A.2d 524, 534-35 *397(describing authoritative control as “right to supervise or control the method of anesthesia, the agents used to achieve the anesthesia, the dosages of those agents,” etc.). This conclusion is further supported by the fact that the hospital did not employ an anesthesiologist and, thus, presumably did not itself have the means of controlling anesthesia decisions.
Also supporting the inference that Dr. Miller had the right to control Nurse Hawkes was plaintiffs evidence that, in an emergency situation, the surgeon has the right to control the anesthetist’s every act. Dr. John B. Neeld, an anesthesiologist, testified that, given the patient’s severe blood loss and precarious life signs, it would have been appropriate for Dr. Miller to have “actively taken over the direction of the anesthesia.” Dr. Robert W. Gaines, an orthopedic surgeon, testified that if he were faced with the sort of emergency encountered by Dr. Miller, he would “enumerate every one of the activities that the CRNA (Certified Registered Nurse Anesthetist) currently should be doing.” Even Dr. Miller admitted that in an emergency situation it is the surgeon who directs the remedial measures taken by the anesthetist. Unlike the Court of Appeals, we do not consider this testimony an incompetent legal conclusion about the scope of a surgeon’s vicarious liability. See 103 N.C. App. at 325-26, 407 S.E.2d at 563. It is testimony regarding the nature of the relationship between surgeon and anesthetist, as defined by the standards of practice prevailing in communities like Beaufort, and, as such, is perfectly competent on the respondeat superior question. See Restatement (Second) of Agency § 220(2)(c).
And, finally, plaintiff supported his claim by showing that Dr. Miller was capable of exercising authoritative control over Nurse Hawkes, in that he knew the principles of anesthesia administration, and that he in fact exercised such control on at least one occasion. At 11:10 a.m., he ordered Nurse Hawkes to stop all anesthesia and give the patient one hundred percent oxygen. Implementing corrective measures in the event of an adverse reaction to anesthesia is one of the functions of the anesthetist. See 21 NCAC 36 .0226 (c)(2)(D). At 11:10 a.m., it was Dr. Miller who decided what corrective action should be taken.
We believe the above evidence reasonably supports the inference that Dr. Miller enjoyed authoritative control over Nurse Hawkes and was, during the surgery, his temporary master. We *398hold, therefore, that the trial court erred in refusing to submit plaintiffs vicarious liability claim to the jury.
 As noted above, the trial court refused to submit the vicarious liability claim to the jury on two grounds. We have established that the first of these — insufficient evidence of a master-servant relationship — was erroneous. We must now evaluate the second: that plaintiff’s release of the servant Hawkes operated, as a matter of law, to release the master Miller.
The trial court’s ruling was based on the common law rule, established in Smith v. R. R., 151 N.C. 479, 66 S.E. 435 (1909), that a release of or covenant not to sue the servant operates to release the master as well. Plaintiff argues correctly that this rule has been abrogated by the Uniform Contribution Among Tort-feasors Act, N.C.G.S. §§ IB-1 to -6 (1983). The Act provides as follows:
When a release or a covenant not to sue ... is given in good faith to one of two or more persons liable in tort for the same injury or the same wrongful death:
1) It does not discharge any of the other tort-feasors from liability for the injury or wrongful death unless its terms so provide ....
N.C.G.S. § IB-4. In the recent case of Yates v. New South Pizza, Ltd., 330 N.C. 790, 412 S.E.2d 666, reh’g denied, 331 N.C. 292, 417 S.E.2d 73 (1992), decided after plaintiff filed his appeal, we interpreted the term “tort-feasors” in the Act to include vicariously liable masters. Thus, the release of a servant no longer operates to release a vicariously liable master, unless the terms of the release so provide. The trial court’s holding was therefore error.
Having rejected both grounds on which the trial court refused to submit plaintiff’s vicarious liability claim to the jury, we now reverse the Court of Appeals and grant plaintiff a new trial on this claim.
 The last issue on appeal relates to plaintiff’s claim that Dr. Miller was negligent in supervising Nurse Hawkes. This claim was submitted to the jury, which answered it in Dr. Miller’s favor. The question is whether the trial court erred in excluding certain *399testimony of plaintiffs expert, Nurse Sandra F. Privatte. Nurse Privatte, qualified as an expert in nurse anesthesia care, would have testified that: 1) Nurse Hawkes needed supervision in ascertaining that there was a medical crisis and in deciding what remedial measures should be taken, and 2) Dr. Miller had a duty to provide such supervision. Plaintiff claims Nurse Privatte was fully qualified to give such testimony and that its exclusion prevented him from proving his negligent supervision claim against Dr. Miller. We agree.
Echoing the trial court, the Court of Appeals held this testimony properly excluded on the ground that Nurse Privatte had not been shown to be familiar with the standard of care for orthopedic surgeons. Under our case law, “the applicable standard of care must be established by other practitioners in the particular field ... or by other expert witnesses equally familiar and competent to testify to that limited field of practice.” Lowery v. Newton, 52 N.C. App. 234, 239, 278 S.E. 566, 571, disc. review denied, 303 N.C. 711, reconsideration denied, 304 N.C. 195, 291 S.E.2d 148 (1981) (emphasis added). Nurse Privatte testified that, in her fifteen years of practice as a nurse anesthetist, she had participated in thousands of operations. Given this experience, she was clearly as knowledgeable as anyone about what a nurse anesthetist can competently do without supervision and what he needs help with. This knowledge, of course, was germane to her own field of practice. However, having worked so frequently with surgeons, she was as knowledgeable as they about the way surgeons ordinarily supervise nurse anesthetists. And, as pointed out by Judge Phillips in his dissent, “what members of a trade or profession ordinarily do in certain situations is evidence of what should be done in those situations.” 103 N.C. App. at 331, 407 S.E.2d at 566. We hold that Nurse Privatte was competent to testify on the matters described and should have been permitted to do so.
The question of whether plaintiff was prejudiced by the improper exclusion of Nurse Privatte’s testimony is a more difficult one. Other of plaintiff’s experts testified that the surgeon is responsible for supervising the remedial measures taken by the anesthetist in a medical crisis. Dr. Robert L. Gibson, an anesthesiologist, testified that, given the bleeding problem, Dr. Miller should have kept himself “constantly aware of all the vital signs of this patient,” and that he was personally responsible for ensuring the patient’s adequate oxygenation. Dr. Neeld also testified that Dr. Miller should have kept himself apprised of the patient’s vital signs and said that *400it was Dr. Miller’s responsibility to decide what remedial measures should be taken: giving more blood, cutting off all anesthetics and giving one hundred percent oxygen, etc. Dr. Gaines testified to the same effect. Thus, it was well established that Dr. Miller had the duty of supervising Nurse Hawkes in the medical crisis. However, only Nurse Privatte offered to testify that Nurse Hawkes in fact needed such supervision and was incapable of making the proper decisions without help. Of all the testimony offered on the issue, this would have been the most persuasive in fixing the high level of supervision required in a medical crisis. Because we believe the admission of this testimony could have altered the jury’s verdict, we now overrule the Court of Appeals and grant plaintiff a new trial on his negligent supervision claim.
Having held that the Court of Appeals erred in affirming the trial court’s directed verdict for Dr. Miller on plaintiff’s vicarious liability claim, we now reverse and remand for a new trial on this claim. Having held that the Court of Appeals erred in affirming the trial court’s exclusion of the testimony of Nurse Privatte, we now reverse and remand for a new trial on plaintiff’s negligent supervision claim. Plaintiff is not entitled to retry his claim that Dr. Miller was negligent in causing, and inadequately treating, the bleeding problem.
REVERSED and REMANDED.
Justice Parker did not participate in the consideration or decision of this case.