There is little, if indeed there is any, need to add to Judge Hincks' opinion in the district court. This being a removed case and depending upon diversity of citizenship, Connecticut decisions are authoritative because the plaintiff was injured in that state. The confusion which has arisen in its disposition arises from that time-honored source of confusion, the maxim, "res ipsa loquitur." The Supreme Court of Connecticut*fn1 has happily accepted Wigmore's gloss,*fn2 and we agree that the circumstances at bar satisfy the conditions there laid down. Furthermore, the gloss in Ryan v. George L. Lilley Co., 121 Conn. 26, 183 A. 2, upon Wigmore's gloss holds, in accordance with the better opinions in other jurisdictions, that although the defendant puts in no evidence to rebut the case made by the maxim, it does not compel the tribunal to find for the plaintiff, although it will support a finding in his favor. Moreover, in the case at bar, even if we were to go still further and assume that the facts compelled a finding that the defendants in the conduct of their work failed in a duty to somebody, the question would remain whether that duty extended to a person standing where the plaintiff stood. Upon that question we must again look to the law of Connecticut,*fn3 which, as Judge Hincks noted, makes the test whether "the ordinary man in the defendant's position, knowing what he knew or should have known" would "anticipate that harm of the general nature of that suffered was likely to result." Incidentally that is not substantially different from the rule as stated in the Restatement of Torts, § 433(b), Comment e. The question upon this appeal therefore becomes whether it was "clearly erroneous" for Judge Hincks not to find that the plaintiff had proved that the defendants should have anticipated, not only that the chain might break, but that a link might be detached and be shot through the air for three hundred feet with speed enough to injure a person standing that far away. We have no hesitation in saying that, not only was that finding not "clearly erroneous," but that any other finding would have been.