R. v. Hemsworth, 2016 ONCA 85:
 This court has repeatedly cautioned against giving undue weight to demeanour evidence because of its fallibility as a predictor of the accuracy of a witness's testimony: Law Society of Upper Canada v. Neinstein, 2010 ONCA 193, 99 O.R. (3d) 1, at para. 66; R. v. Rhayel, 2015 ONCA 377, 324 C.C.C. (3d) 362. As I indicated in Rhayel, at para. 85, "[i]t is now acknowledged that demeanour is of limited value because it can be affected by many factors including the culture of the witness, stereotypical attitudes, and the artificiality of and pressures associated with a courtroom."
 Although the law is well settled that a trial judge is entitled to consider demeanour in assessing the credibility of witnesses, reliance on demeanour must be approached cautiously: see R. v. S. (N.), 2012 SCC 72,  3 S.C.R. 726, at paras. 18 and 26. Of significance in this case is the further principle that a witness's demeanour cannot become the exclusive determinant of his or her credibility or of the reliability of his or her evidence: R. v. A. (A.), 2015 ONCA 558, 327 C.C.C. (3d) 377, at para. 131; R. v. Norman (1993), 16 O.R. (3d) 295 (C.A.), at pp. 313-14.
The Trial Judge's Reliance on the Appellant's Demeanour
 It is of note that the trial judge started his assessment of the appellant's credibility by expressing his concern about how the appellant testified:
It was my impression that the [appellant] testified in a careful fashion which lacked any spontaneity. He appeared to me to be a witness who was prepared and aware of what his evidence should be to raise a reasonable doubt as opposed to a retired teacher wrongfully accused of fondling a young male 20 years ago.
 This paragraph troubles me for two reasons.
 First, the trial judge found the appellant's credibility was diminished because he testified in a manner that was too careful. The problem I have with this is that the trial judge had no reference point. He did not know how the appellant normally expressed himself. Moreover, the appellant's "careful fashion" of testifying may relate to such factors as the unfamiliar atmosphere of the courtroom, the artificiality of the circumstances under which the appellant was being asked to provide information and the pressure he was under given what was at stake.
 Second, I am concerned by the trial judge's finding that the appellant testified more like someone trying to raise a reasonable doubt than someone wrongfully accused of sexually assaulting a student. The trial judge provided no insight as to how he arrived at that conclusion, particularly how the differences between the two types of testimony might manifest themselves. General assertions such as this are, with respect, not only unhelpful but also defy appellate review.
 Although the trial judge was entitled to consider the appellant's demeanour in assessing his credibility, he erred by considering factors that had little, if any, evidentiary foundation, and by generally over-emphasizing how the appellant appeared in the witness box.
 Again, the error must be material. In my view, it is. It must be remembered that the trial judge's assessment of the appellant's demeanour was one of only two factors the trial judge relied on in rejecting the appellant's evidence. Against that background, and given the singular importance of the appellant's testimony, the errors the trial judge made in factoring the appellant's demeanour into the trial judge's assessment of the appellant's credibility were material.
Of the Law Societies of Upper Canada and Nunavut