During a steam line renovation project at the campus of the University of Massachusetts in Amherst, Paul Hammell, Jr., a worker at the project and the plaintiff here, suffered severe burns from a burst of high-pressure steam through a ruptured pipeline. He brought a personal injury action and tried the case to a jury in the Superior Court against four defendants: Shooshanian Engineering Associates, Inc. (Shooshanian), the project’s designer; All State Construction Services, Inc. (All State), the general contractor; the University of Massachusetts (university); and the University of Massachusetts Building Authority (building authority), which had hired Shooshanian and All State and funded the project. The jury found the university and Shooshanian negligent, concluded that only the university’s negligence had caused the injury, and awarded Hammell $392,397.48. The judge reduced the amount to $100,000.00 in compliance with the damages cap imposed by the Massachusetts Tort Claims Act, G. L. c. 258, § 2.

The trial judge’s rulings upon several motions have generated this appeal. The university filed a motion for a directed verdict at the conclusion of Hammell’s case; renewed it at the close of all evidence; and moved for a judgment notwithstanding the verdict. Hammell filed a motion for a new trial against Shooshanian and All State on the ground that the testimony of an expert witness for Shooshanian was prejudicial because it deviated significantly from the expert’s earlier disclosures on central factual issues.2 The university filed a motion for a new trial also on that ground. The university and Hammell appeal from the denial of those motions. We affirm the judge’s denials of the motions appealed from, and affirm the judgments on the jury verdicts.

1. Background. The university owns and operates a steam generating plant supplying heat and hot water to facilities across its Amherst campus through a network of underground steam *636pipes. The building authority hired Shooshanian to design, and All State to coordinate and implement, the replacement of aging pipes. The university was responsible for turning the steam on and off as necessary to accommodate the project.

Water in a high-pressure steam pipe is a hazard. The pressure can propel the water particles to a very high velocity, a phenomenon called a “water hammer.” A water hammer can gather enough force to rupture the pipe on contact. Steam condensation results in water in the pipes. Prevention of water hammers requires the slow start-up of transmission of steam through the pipes. Reactivation of steam should commence only after residual water has had sufficient time to drain; the pressure must increase gradually to allow for drainage of any new condensate before the steam achieves a high velocity.

The accident in question occurred on campus in “manhole C.” Manhole C contained a high-pressure steam pipe leading to an area called Garber Field, and a low-pressure steam pipe leading to a building known as Boyden Gymnasium. The design specifications by Shooshanian called for the installation of a pressure-reducing valve linking the two pipes. The specifications otherwise left the high-pressure pipe untouched. All State, however, cut into the high-pressure pipe and formed a new end with a cap to interrupt the flow of steam to Garber Field. It did so out of concern for the safety of workers excavating the area.

When All State had completed the work in the manhole, the university reactivated the high-pressure steam pipe. On July 30, 1997, at the direction of his supervisor, Hammell entered manhole C and opened a valve enabling the steam to flow through the newly installed pressure-reducing device from the high-pressure pipe to the low-pressure pipe. He heard two loud bangs. The high-pressure pipe suddenly released scalding steam into the manhole. Hammell suffered severe bums.

2. Discussion, a. Motions for a new trial. By answer to interrogatories nine months before trial, Shooshanian had represented to all parties that its expert witness, registered professional engineer David Elovitz, would testify that, when Hammell opened the valve, a water hammer hit the capped end of the high-pressure pipe with great force; that the excess water in the pipe had resulted from inadequate drainage caused by hasty buildup *637of steam pressure during a warm-up period of less than one hour. He would opine that, if All State had followed Shooshanian’s specifications, (1) the pipe would have remained uncut, uncapped, and connected to a rigid structure; (2) excess water would have drained more effectively before and during the activation of steam; (3) a water hammer, if any, would have exerted less force on the pipe because it would not have hit the capped end; and (4) the pipe would have absorbed the force of a water hammer without slamming into a steel structure on the manhole floor and rupturing.

On the one hand, Elovitz’s expected testimony implicated the university for reactivating the steam pressure in the pipe too quickly and causing the water hammer. On the other hand, Elovitz’s opinion faulted All State for deviation from Shooshanian’s specifications so as to cause increased condensate, a more forceful water hammer, and a pipe weakened by cutting and capping and unable to withstand the impact of the water hammer.

To accommodate Elovitz’s schedule, the judge permitted him to testify during Hammell’s case-in-chief. The judge ruled that, for the purpose of a motion for a directed verdict, he would not consider Elovitz’s testimony as part of Hammell’s case. At trial, without prior notice to the other parties, and over timely motions by Hammell and the university to strike the testimony from the record, Elovitz stated for the first time that (1) under the contract, All State had the discretion to cut and cap the high-pressure pipe; (2) the force of the water hammer was so great that the pipe would have hit the steel structure even if it were not cut and capped; and (3) the overly fast activation of steam pressure, and not the cutting and capping, had caused the water hammer and the rupture of the pipe. He explained that two weeks before trial he had realized that Hammell had opened a valve different from the one which he had assumed originally. That fact prompted him to change his opinion.3 Essentially, his revised opinion faulted the university and absolved All State and Shooshanian.

*638The university and Hammell argue that the unexpected change of Elovitz’s opinion caused prejudice warranting a new trial. The admissibility of belatedly disclosed expert opinion rests within the sound discretion of the trial judge. Resendes v. Boston Edison Co., 38 Mass. App. Ct. 344, 350 (1995). We review the judge’s denial of the new trial motions for abuse of discretion. Ibid.

The extreme sanction of a new trial requires both surprise and unfair prejudicial harm. Beaupre v. Cliff Smith & Assocs., 50 Mass. App. Ct. 480, 485-486 (2000), citing Resendes, supra at 351. Several factors work against the finding of such unfair prejudice here: neither Hammell nor the university requested a continuance; both conducted extensive and effective cross-examination of Elovitz; and both were on notice for over a year before the trial that the timing of the activation of steam was at issue. See id. at 486, and cases cited; Elias v. Suran, 35 Mass. App. Ct. 7, 10-11 (1993). See also Newell Puerto Rico, Ltd. v. Rubbermaid Inc., 20 F.3d 15, 20-22 (1st Cir. 1994).

The last factor is substantial. Sixteen months before trial, Hammell had disclosed that his expert, professional engineer Alden R Gaudreau, would testify that the university did not reactivate the high-pressure pipe “in a safe and proper manner prior to the explosion.” While Gaudreau did not testify, Hammell listed him as an expert witness in a joint pretrial memorandum and reiterated his opinion. Similarly, All State had reported in its supplemental answer to interrogatories that its expert, professional engineer Joseph Grynbaum, would opine that All State had complied with “good and accepted construction practices”; that the *639university had failed to anchor properly the high-pressure pipe; and that the university “was responsible for re-energizing -the steam [pipes] in a safe and reasonable manner.” This opinion also put at issue the university’s reactivation of steam in the allegedly weakened pipe. Although Grynbaum did not testify, in the joint pretrial memorandum All State had alleged similarly that the university “was responsible for re-energizing the steam [pipes] in a safe and reasonable manner” and that it “did not take the proper precautions to ensure that the high pressure steam [pipe] had completely re-energized.” Notable also is Elovitz’s original opinion, which expressly called into question the activation of steam in less than an hour. Those pretrial disclosures warned the university of the need to defend the process of its reactivation of the pipe. In the joint pretrial memorandum submitted approximately eight months before trial, it did not identify an expert witness of its own, nor did it call one at trial.

Those same pretrial sources show Hammell’s knowledge of the need to present evidence at trial of both the university’s overly rapid reactivation of the steam line and All State’s allegedly imprudent cutting and capping of the manhole pipe. By multiple means, he indicated his awareness that he could not rely upon the interrogatory answer of Elovitz for the essential element of his prima facie claim against All State of concurrent proximate causation of the accident by means of its cut-and-cap treatment of the manhole pipe. He identified his own expert (Gaudreau) as a witness expected to opine that All State’s action allowed the pipe to destabilize and the steam to escape. He was aware from answers to interrogatories and the pretrial memorandum that All State’s expert (Grynbaum) would characterize that proximate cause as “speculative.” As part of his case-in-chief, Hammell called professional engineer Michael Carr, a former employee of Shooshanian, to testify to his inspection of the manhole one day after the accident and to his impression that the cause of the escape of the steam was “that the pipe was incorrectly cut and capped.” Hammell did not identify Elovitz (and therefore Elovitz’s original opinion) as an element of his case-in-chief. His counsel vigorously cross-examined Elovitz about his reversal of opinion and exposed it to the jury. In these circumstances, the trial judge could reasonably conclude that Elovitz’s revision did *640not cause genuinely harmful surprise to the university or Hammell and that neither party was entitled to a new trial upon that ground.

b. Sufficiency of the evidence of liability against the university. The university submitted a motion for a directed verdict at the conclusion of the plaintiff’s evidence; a motion for a directed verdict at the conclusion of all the evidence; and a motion for judgment notwithstanding the verdict after trial. See Mass.R.Civ.P. 50(a) & (b), as amended, 428 Mass. 1402 (1998). It appeals exclusively from the denial of the first motion for a directed verdict. It cannot do so because its subsequent submission of evidence as a defendant effectively waived its right to challenge the denial of the earlier motion. The reason for the rule is that, once the defendant has added to the fund of evidence, the judge and the jury “have the right to consider the whole case” upon the totality of the testimony and exhibits accumulated through the close of evidence. Martin v. Hall, 369 Mass. 882, 884-885 (1976), quoting from Bogk v. Gassert, 149 U.S. 17, 23 (1893). Evidence from a defendant may fill gaps in a plaintiff’s prima facie case. That probability increases in circumstances, as here, in which multiple defendants may attempt to prove liability against each other. For purposes of appeal, a defendant therefore cannot retrospectively freeze the evidence into its state at the conclusion of the plaintiff’s case. See Martin, supra-, McMahon v. Finlayson, 36 Mass. App. Ct. 371, 374 n.6 (1994). See also Smith & Zobel, Rules Practice § 50.11 (2d ed. 2007). As a result of the university’s effective waiver of appeal from the rulings upon its three motions, we need not measure the sufficiency of Hammell’s evidence of liability against it.4

Judgments affirmed.