Evans, J.
1- ERRonf subsetewnofaease!: The case was before us upon a former appeal. 181 Iowa 65. Upon that appeal, we reversed a judgment of the lower court in favor of the plaintiff. For a full discussion of the case, reference may be had to the opinion ,on that appeal. The general ground of the reversal was that the defective sidewalk complained of 'by the plaintiff was not such as could be said to render the defendant liable as for actionable negligence. It is now claimed that additional evidence was introduced upon the second trial which completely obviates and avoids the opinion of this court, and that the evidence in its present condition would sustain a verdict for the plaintiff. The case has been argued de novo by the plaintiff, and it is earnestly contended that the trial court erred in directing a verdict.
Needless to say that the opinion on the first appeal has become the law of the case. The main question for our consideration is whether the evidence on the second trial *62can be said to be materially different from that upon the first. The accident occurred on July 28, 1914, in the daytime. It occurred at a place where there was a defect in the sidewalk. The defect consisted in the deterioration, to a greater or less' degree, of a section of a cement sidewalk, about 4 feet in length. This deterioration caused a depression in the level of the walk. The claim of plaintiff is that, while walking in a procession over said walk, and stepping forward into the alleged depression, she fell to her knees; and that such fall resulted in greatly injuring her internal organs; and that the damages resulting are very great.
At the first trial, the nature and extent of the defect in the sidewalk was gone into fully. Plaintiff introduced several witnesses who testified to their estimates of the extent of deterioration, and of the depth of the depression. Photographs had been taken, and three of these were introduced in evidence by the plaintiff. The same photographs have been introduced in evidence on the second trial. At the former trial, plaintiff’s witnesses estimated the extreme depth of the depression in the defective section to bo from 2 to 3 inches. At the second trial, the plaintiff pro duced additional witnesses, some of whom estimated the depth at from 3% to 4 inches. ■ These were estimates only, and based wholly upon the present recollection of observations made more than three years prior. The estimates given by the additional witnesses were as follows: Downey, “3 inches;’’ Horsely, “3 to 3% inches;” Roll, “3 to 4 inches;” Mrs. Roll, “perhaps 3% inches;” Coats, “at that time at the west end the grouting was out clear to the cinders. The top coat in the grouting was about 3y2 or 4 inches ’ thick.”
*632. Appeal and error : second appeal: nature and sufficiency of additional evidence. *62The contention for the plaintiff is that the evidence as now presented shows a depth of the depression of at least 4 inches, and that the case should be considered anew *63in the light of such additional testimony. Naturally, the plaintiff contends that she was entitled at her second trial to try her case de novo. All this being conceded, it does not follow that the plaintiff may find herself unhampered at a second trial with the testimony put forward by her upon the first. The former evidence is still in the record, and the same photographs were used upon both trials. The question confronting us now is whether it can fairly be said that the evidence upon this trial is so substantially different as to obviate our holding on the first appeal. In passing upon this question, we must take account .of the nature of the additional evidence which is offered, it is opinion evidence; a mere estimate; and an estimate based upon mere recollection .of observations had a long time ago. An estimate at best is an approximation. Where it is based wholly upon the recollection of observations a long time past, the range .of its uncertainty is necessarily increased. It becomes doubly elastic, and, like a rubber band, may readily take up an inch or give one. We do not hereby impute any bad faith to any of this testimony. On the contrary, we accept it as given in good faith. But, in considering the question before us, we ought not to close our eyes to the nature of this class of testimony and to the infirmity that inheres in it. Boeck v. Modern Woodmen, 183 Iowa 211. Taking the record as presented to us on the former appeal, and comparing therewith the record as now presented, including the additional evidence descriptive of the alleged defect in the sidewalk, we think it must fairly be said that no substantial distinction is presented. The ruling of the trial court to that effect was proper, and its judgment must be — Affirmed.
Laud, C. J., Preston and Salinger, JJ., concur.