The plaintiff, Robert Douillard, appeals from the allowance of the defendant’s motion for summary judgment, in which a judge in the Superior Court held that the plaintiff’s evidence of negligent service of alcohol was insufficient to raise a genuine issue of material fact. The plaintiff appealed and we granted his application for direct appellate review. We now *163vacate the summary judgment and remand the matter to the Superior Court for trial.
1. Background. Viewed in the light most favorable to the plaintiff, and drawing all permissible inferences in his favor, the record on summary judgment demonstrates that the defendant LMR, Inc., operated a tavern known as Breaker’s Billiards & Brews. On the evening of March 25, 1994, Steven Gagne was a patron at the defendant’s establishment, arriving at approximately 8 p.m. and remaining there until sometime after midnight. Gagne had not consumed any alcoholic beverages that evening prior to his arrival at the defendant’s tavern. He had eaten three meals that day. Gagne claimed that, during the roughly four hours he was at the bar, he drank four drinks of rum and coke, that he was served a fifth drink, and that he had consumed a portion of that fifth drink before his departure. Gagne is five feet two and one-half inches tall and, at the time of these events, weighed between 130 and 140 pounds.
At around 1 a.m. on March 26, 1994, Gagne’s vehicle crossed into the oncoming lane and struck a vehicle operated by the plaintiff, Robert Douillard. Both Douillard and Gagne were seriously injured in the collision. When examined at a hospital, a blood sample drawn from Gagne revealed that he had a blood alcohol concentration of .149 per cent as of 1:56 a.m. For his role in this incident, Gagne was charged and convicted of operating while under the influence resulting in serious bodily injury and operating negligently so as to endanger.
Four of Gagne’s friends and acquaintances testified in Gagne’s defense at his criminal trial. They had seen Gagne at the bar that night and testified that, in all respects up through the time of his departure, he had shown no signs of intoxication whatsoever. At deposition, one of those same friends testified that he had, on approximately ten prior occasions, seen Gagne when he was intoxicated. The friend described that, when intoxicated, Gagne became “[ojverly social, loud, sick, [and] giddy.” The witness further specified that by the term “overly social” he meant “[hjugging people, laughing” and “[l]oud behavior,” that by the term “sick” he meant “[v]omiting,” and that by the term “giddy” he meant “laughing inappropriately.” At his deposition, Gagne testified that he had, on prior occasions, become intoxicated after drinking seven rum and cokes *164and that, in his experience, it took seven such drinks to render him intoxicated.
In opposition to the summary judgment motion, the plaintiff submitted an affidavit from an expert witness setting forth a calculation of Gagne’s blood alcohol level at the time he was served his last drink. Based on the time frames testified to by Gagne, his height and weight, his consumption of food that day, and his blood alcohol level as determined by the hospital at 1:56 a.m., the expert opined that Gagne’s blood alcohol level would have been .154 per cent at the time he was served his last drink. He then opined as follows:
“My education, experience and training has clearly demonstrated to me that an individual who has a blood alcohol level of .12% or higher clearly demonstrates signs of intoxication. While all these signs and symptoms are not present in every individual on every drinking occasion, at least some of them are. These include a fixating of the eyes, the so-called ‘glassy eye.’ The individual begins to have trouble with visual focus and will tend to fixate on an object for a prolonged period of time. Individuals also can become more talkative and their speech patterns and tone begin to change. This normally exhibits itself by an increase in both pitch and volume of speech. Some individuals also demonstrate difficulty with balance and coordination. . . . It is my opinion to a reasonable degree of scientific certainty that Mr. Gagne was demonstrating obvious signs of intoxication when he was served his last drink at the defendant’s establishment on the evening of March 25, 1994.”
The expert also opined that, to reach the blood alcohol level shown at the hospital at 1:56 a.m., Gagne would have had to consume at least nine drinks that evening, not the four and one-half drinks that Gagne had acknowledged.
2. Obvious intoxication. “[A] tavern keeper does not owe a duty to refuse to serve liquor to an intoxicated patron unless the tavern keeper knows or reasonably should have known that the patron is intoxicated.” Vickowski v. Polish Am. Citizens Club of Deerfield, Inc., 422 Mass. 606, 609 (1996), quoting Cimino v. Milford Keg, Inc., 385 Mass. 323, 327 (1982). Thus, on a claim for negligent service of alcohol to an intoxicated patron, a plaintiff must come forward with some evidence that the *165patron’s intoxication was apparent at the time he was served by the defendant. See Vickowski v. Polish Am. Citizens Club of Deerfield, Inc., supra at 608-609, 610; Cimino v. Milford Keg, Inc., supra at 327-329; Makynen v. Mustakangas, 39 Mass. App. Ct. 309, 314 (1993); Hopping v. Whirlaway, Inc., 37 Mass. App. Ct. 121, 124 (1994); Kirby v. Le Disco, Inc., 34 Mass. App. Ct. 630, 631-632 (1993); Wiska v. St. Stanislaus Social Club, Inc., 7 Mass. App. Ct. 813, 816-818 (1979).
While our case law has identified apparent intoxication at the time of service as an element that must be proved on such a claim, nothing imposes on the plaintiff any higher or special burden of proof with respect to this element and, as in any other case, the plaintiff may seek to prove that element by direct evidence, circumstantial evidence, or a combination of the two. See Vickowski v. Polish Am. Citizens Club of Deerfield, Inc., supra at 611 (direct evidence of obvious intoxication not required). Thus, the only burden on the plaintiff here is to show that it is more probable than not that Gagne appeared intoxicated at the time he was served his last rum and coke. He does not have to establish this element to any greater degree of certainty. And, as in any other case, if circumstantial evidence, including expert testimony, would permit the jury to infer that Gagne probably appeared intoxicated at the time he was served, the plaintiff’s claim should survive summary judgment.
Evidence of apparent intoxication, or of elevated blood alcohol levels, at some later point in time does not, by itself, suffice to show that the patron’s intoxication was evident at the time the last drink was served. See McGuiggan v. New England Tel. & Tel. Co., 398 Mass. 152, 162 & n.10 (1986) (expert opinion as to visible intoxication at time guest left party does not relate to time at which guest was last served alcohol). “Our reluctance to accept such evidence as sufficient stems from the uncertainties of the situation, including the possible delayed impact .of the consumption of alcohol, and the unknown effect on a patron of the last drink served to him by a licensee.” Vickowski v. Polish Am. Citizens Club of Deerfield, Inc., supra at 612. See Makynen v. Mustakangas, supra at 313. From the mere fact of intoxication observed at some later time (e.g., at the accident scene), one could not tell what contribution the patron’s final drink had made toward that state of intoxication. Evidence of later intoxication has been admitted for purposes of bolstering other evidence concerning a patron’s condition at the *166time alcohol was served, Vickowski v. Polish Am. Citizens Club of Deerfield, Inc., supra at 612 n.5, citing Gottlin v. Graves, 40 Mass. App. Ct. 155, 159 (1996), but it is not sufficient by itself to establish apparent intoxication at the time that alcohol was served.
In the present case, the plaintiff has submitted expert testimony to provide the necessary link between an elevated blood alcohol level some two hours after Gagne’s last drink and signs of intoxication at the time he was served that last drink. Specifically, the expert has calculated that Gagne’s blood alcohol level was at .154 per cent as of the time he was last served, using a methodology that the defendant has not yet challenged in any way.1 The expert then opines that one would normally see signs of intoxication at blood alcohol levels of only .12 per cent, a reading well below Gagne’s. Based on this expert analysis, the plaintiff contends that, if most people show some sign of intoxication at a blood alcohol level of .12 per cent, a rational jury could find it more probable than not that Gagne was showing signs of intoxication at his blood alcohol level of .154 per cent, the level he had reached by the time he was last served a drink at the defendant’s tavern.
The use of such expert testimony relating blood alcohol level to the appearance of signs of intoxication has been rejected by some courts on the ground that it fails to take into account the wide individual variability in alcohol tolerance. See Romano v. Stanley, 90 N.Y.2d 444, 450-451 (1997); Johnson v. Harris, 419 Pa. Super. 541, 551-552 (1992); Purchase v. Meyer, 108 Wash. 2d 220, 224-227 (1987). It is common knowledge that some people are, in common parlance, better able to “hold their liquor” than others. Thus, it is argued, expert opinion concerning what signs of intoxication are displayed by the “average” person, or by “most” people, at particular blood alcohol levels should not suffice to prove whether a particular individual actually displayed such signs at those levels.2
Assuming, without deciding, that the problem of individual *167variability in response to alcohol prevents a plaintiff from relying exclusively on expert opinion to make out a case of apparent intoxication, the present case does not rely solely on the expert’s opinion about what signs of intoxication a normal, average, or typical drinker would exhibit at a blood alcohol level of .154 per cent. This record adds to the expert opinion specific information concerning this particular drinker’s reaction to alcohol consumption. Gagne himself acknowledged that he would, after consuming seven drinks, become intoxicated. Where the expert calculated that Gagne’s consumption on the night in question had to be at least nine drinks, there is evidence that the amount consumed by Gagne put him at least two drinks beyond the level of alcohol consumption at which he would customarily become intoxicated. As to whether Gagne, when intoxicated, was a person whose intoxication would manifest outward signs, a friend who had seen him intoxicated on ten occasions confirmed that Gagne would indeed exhibit such signs. According to Gagne’s friend, excessive alcohol consumption would make Gagne “[o]verly social, loud, sick, [and] giddy.” When intoxicated, Gagne would “[h]ug[] people,” engage in “[l]oud behavior,” “[v]omit,” and “laugh[] inappropriately.” In other words, where the expert opinion was that one would expect to see signs of intoxication at the blood alcohol level that Gagne reached after his estimated nine drinks, the testimony of Gagne and his friend confirmed that Gagne would customarily become intoxicated and exhibit several signs of intoxication after consuming only seven drinks.
Thus, the plaintiff’s case does not rest solely on an expert’s assessment of how the “average” drinker would react to this much alcohol. Rather, there is direct evidence of Gagne’s own reactions to alcohol to confirm that they are in fact comparable to the average drinker. Where the record contains such specific information about Gagne’s own reactions to excessive consumption, we do not need to decide whether expert testimony about the average drinker’s response to alcohol would be sufficient to *168demonstrate that a particular drinker more probably than not appeared intoxicated.3
Where the plaintiff’s opposition to the summary judgment motion included affidavits and deposition testimony demonstrating that plaintiff had sufficient evidence to permit the inference that Gagne was visibly intoxicated at the time he was served his final drink, it was error to grant the defendant’s motion for summary judgment. The judgment is therefore vacated and the matter is remanded to the Superior Court for further proceedings consistent with this opinion.
So ordered.
Justice Abrams participated in the deliberation on this case, but retired before the opinion was issued.
The fact that the defendant’s summary judgment motion did not contest the expert’s qualifications or methodology does not preclude the defendant from making such challenges at trial.
In McGuiggan v. New England Tel. & Tel. Co., 398 Mass. 152, 155, 162 & n.10 (1986), we assumed without deciding that such a “generalized” opinion would he admissible. Our concerns with that expert opinion were based on the expert’s failure to address “specific circumstances” such as time of the alcohol consumption and food consumption. The expert here has considered *167the timing of Gagne’s drinks and the fact that he had eaten three meals that day. It also relates to the relevant time period, i.e., to the time at which Gagne was served his last drink. As such, it is admissible. The expert may of course be subject to cross-examination, and we express no opinion as to the validity of the expert’s methodology or calculations.
The plaintiff also contends that the expert’s opinion with respect to the consumption of nine drinks in the space of four hours would, by itself, allow the fact finder to conclude that Gagne’s intoxication had become apparent by the time of his ninth drink (if not before). “[A] jury confronted with evidence of a patron’s excessive consumption of alcohol, properly could infer, on the basis of common sense and experience, that the patron would have displayed obvious outward signs of intoxication while continuing to receive service from the licensee.” Vickowski v. Polish Am. Citizens Club of Deerfield, Inc., 422 Mass. 606, 611 (1996). The defendant argues that the plaintiff “waived” any “theory of excessive consumption” by not raising the issue in his memorandum in opposition to summary judgment. That' excessive consumption can support the inference of apparent intoxication is not a separate “theory” of opposition. It is merely one form of circumstantial evidence which may be used (along with others) to support the inference that a drinker’s intoxication would likely have been apparent. Id. Because we have determined that the expert’s affidavit and the direct evidence of Gagne’s customary reaction to alcohol sufficed to defeat the defendant’s summary judgment motion, we need not address the argument that the quantity of alcohol consumed was such that the jury could infer apparent intoxication.