Evidence - Circumstantial Evidence. R. v. Wood
In R v Wood (Ont CA, 2022) the Court of Appeal commented succinctly on the nature of circumstantial evidence [Watt JA]:
 The final point concerns circumstantial evidence. Like any item of evidence, circumstantial evidence must be relevant, material, compliant with any applicable admissibility rule and more probative than prejudicial. Circumstantial evidence is all about inferences. An inference is a deduction of fact that may logically and reasonably be drawn from another fact or group of facts found or otherwise established in the proceedings. Inferences arise from objective facts that give rise to those inferences. Absent such facts from which the inferences may be drawn, no inference arises, only impermissible speculation and conjecture.. R v Gibson
In R v Gibson (Ont CA, 2021) the Court of Appeal considered several points about circumstantial evidence in a criminal appeal:
 First, circumstantial evidence is all about inferences. Individual items of circumstantial evidence give rise to a range of inferences. The available inferences must be reasonable according to the measuring stick of human experience. That there may be a range of inferences available from an individual item of circumstantial evidence does not render the item of evidence irrelevant or neutralize its probative value: R. v. Calnen, 2019 SCC 6,  1 S.C.R. 301, at para. 112, per Martin J. (dissenting, but not on this point), citing R. v. Smith, 2016 ONCA 26, 333 C.C.C. (3d) 534, at para. 77.. R. v. Stennett
 A second point concerns the standard of proof required where proof of the offence or one or more of its essential elements depends wholly or substantially on circumstantial evidence. In such a case, an inference of guilt drawn from circumstantial evidence must be the only reasonable inference available on that evidence: R. v. Villaroman, 2016 SCC 33, at paras. 30, 32-34.
 Third, the standard of proof applies to the evidence taken as a whole, not to each individual item of circumstantial evidence: R. v. Morin, 1988 CanLII 8 (SCC),  2 S.C.R. 345, at pp. 359, 362. See also R. v. Morin, 1992 CanLII 40 (SCC),  3 S.C.R. 286, at p. 295-96.
 Finally, where proof of an essential element or the offence charged depends wholly or substantially on circumstantial evidence, it is the cumulative effect of all the evidence, taken together, each item in relation to another and the whole, that must be considered in determining whether the standard of proof has been met: Cote v. The King (1941), 1941 CanLII 348 (SCC), 77 C.C.C. 75 (S.C.C.), at p. 76.
In R. v. Stennett (Ont CA, 2021) the Court of Appeal (Watt JA) considered the requirements of circumstantial evidence:
 Fourth, circumstantial evidence and the burden and standard of proof.. R. v. Aslami
 Circumstantial evidence is about drawing inferences and the range of reasonable inferences that may be drawn from that evidence. In an assessment of circumstantial evidence, a trier of fact is required to consider, or be directed to consider, “other plausible theories” and “other reasonable possibilities” inconsistent with guilt. In its effort to establish guilt, the Crown must negate those “reasonable possibilities”. However, the Crown need not negative every possible conjecture, no matter how irrational or fanciful, which might be consistent with the innocence of the accused. These “other plausible theories” or “other reasonable possibilities” must be rooted in logic and experience applied to the evidence or absence of evidence: R. v. Villaroman, 2016 SCC 33,  1 S.C.R. 1000, at para. 37. See also R. v. Bagshaw, 1971 CanLII 13 (SCC),  S.C.R. 2, at p. 8.
 Circumstantial evidence does not have to totally exclude other conceivable inferences. The trier of fact must not act on alternative interpretations of the circumstances that it considers unreasonable. Alternative inferences, whether arising from the evidence or an absence of evidence, must be reasonable, not simply possible: Villaroman, at para. 42. Non-culpable inferences, based on the evidence or the absence of evidence, must be reasonable when assessed logically and in light of human experience and common sense: Villaroman, at para. 36.
In R. v. Aslami (Ont CA, 2021) the Court of Appeal commented on the requirements of circumstantial evidence in a criminal case:
 As I noted at the outset, this was an entirely circumstantial case. Consequently, the trial judge was required to follow the analytical route set out in Villaroman, where Cromwell J. said, at para. 37:
When assessing circumstantial evidence, the trier of fact should consider "other plausible theor[ies]" and "other reasonable possibilities" which are inconsistent with guilt. I agree with the appellant that the Crown thus may need to negative these reasonable possibilities, but certainly does not need to "negative every possible conjecture, no matter how irrational or fanciful, which might be consistent with the innocence of the accused". "Other plausible theories" or "other reasonable possibilities" must be based on logic and experience applied to the evidence or the absence of evidence, not on speculation. [Citations omitted.] [Emphasis in original.] The trial judge purported to comply with this approach, but it is clear that he did not do so. He said that he could see “no other ‘plausible theories/reasonable possibilities’ other than that of the guilt of [the appellant.]” It seems likely that it was the errors, to which I have referred above, that caused him to reach that conclusion.
 It also seems likely that that conclusion was the result of the trial judge not considering the evidence as a whole. Viewing the evidence as a whole is a critical part of its necessary evaluation, especially in a circumstantial case, in determining whether it proves guilt beyond a reasonable doubt. As Watt J.A. said in R. v. Smith, 2016 ONCA 25, 333 C.C.C. (3d) 534, at para. 81: “It is essential to keep in mind that it is the cumulative effect of all the evidence that must satisfy the criminal standard of proof, not each individual item which is merely a link in the chain of proof” (citations omitted).