Mr. Justice Eakin
delivered the opinion of the court.
1. The plaintiff’s case rests principally upon whether the statements made by Smith, the engineer, to witnesses Olsen and Mahoney, soon after the accident, in relation to the cause thereof, was competent evidence against the principal as admissions by an agent. It was not shown by the evidence that Smith was the agent of the defendant for any. purpose except to operate the winch and derrick. His acts within that employment are the acts *37of the principal, and what he said about those acts while performing them is a part of the transaction; and, where the acts are competent evidence against the principal, such statements are also admissible as characterizing the acts and constituting a part of them; but statements of a past transaction made by an agent, and not part of the acts itself or characterizing it, are not within the scope of the agency, and cannot be admitted in evidence to affect the principal: First National Bank v. Linn County Bank, 30 Or. 296 (47 Pac. 614) ; Wicktorwitz v. Farmers Ins. Co. 31 Or. 569 (51 Pac. 75) ; Luman v. Golden A. C. M. Co. 140 Cal. 700 (74 Pac. 307); Redmon v. Metropolitan St. Ry. Co. 185 Mo. 1 (84 S. W. 26: 105 Am. St. Rep. 558).
2. In Vicksburg & M. R. Co. v. O’Brien, 119 U. S. 99 (7 Sup. Ct. 172: 30 L. Ed. 299), Mr. Justice Harlan, in considering the admissibility of declarations of an agent as against the principal, says:
“There can be no dispute as to the general rules governing the admissibility of the declarations of an agent to affect the principal. * * The admission or declaration of his agent binds him only when it is made during the continuance of the agency in regard to a transaction then depending, ‘et dum fervet opus.’ It is because it is a verbal act and part of the res gestae, that it is admissible at all. and therefore it is not necessary to call the agent to prove it; but, wherever what he did is admissible in evidence, there it is competent to prove what he said about the act while he was doing it. * * But an act done by an agent cannot be varied, qualified, or explained, either by his declarations, which amount to no more than a mere narrative of a past occurrence, or by an isolated conversation held, or an isolated act done, at a later period.”
The statements of Smith sought to be proved here were no part of his acts in operating the winch, but only his account or opinion of it given afterward in response to an inquiry, and were incompetent to affect the defendants.
*383. It is also urged by plaintiff that the statements of Smith referred to were admissible as a part of the res gestae of the accident because made soon thereafter. Some of the courts have permitted a wide range in the admission of statements made after the principal act, as part of the res gestae, as being largely in the discretion of the trial court, on the theory that because of its proximity in time to the main act or transaction there is a strong probability that the declarations are true. In Vicksburg & M. R. Co. v. O’Brien, 119 U. S. 99 (7 Sup. Ct. 172: 30 L. Ed. 299), Mr. Justice Harlan, for the majority of the court, as to such statements being res .gestae, says: “The cases have gone far enough in the admission of subsequent declarations of agents, as evidence against their principals”—while Mr. Justice Field, dissenting, says: “The modern doctrine has relaxed the ancient rule that declarations, to be admissible as part of the res gestae, must be strictly contemporaneous with the main transaction.” However, it seems unnecessary to look beyond the decisions of our own court for the solution of this question. Section 698, B. & C. Comp., provides that, “Where also the declaration, act, or omission forms part of a transaction which is itself the fact in dispute, or evidence of that fact, such declaration, act, or omission is evidence as part of the transaction”; thus defining res gestae. In Sullivan v. O. R. & N. Co., 12 Or. 392, 398 (7 Pac. 508, 512: 53 Am. Rep. 364) j in discussing this question, Mr. Justice Thayer says:
“The Massachusetts cases, with the exception of the one referred to, have generally held to a reasonable and consistent rule upon that branch of evidence. They have repudiated the notion that the admission of such declarations is left to the discretion of the presiding judge, and admit them only when they are calculated to explain the character and quality of the act, and are so connected with it as to derive credit from the act itself, and to constitute one transaction. This appears to me *39to be as liberal a rule as any court can, consistently with the rules of evidence, sanction, and I think it very doubtful whether our courts, under certain provisions of our statute, would have any right to permit the introduction of declarations of parties as evidence, except under the condition of circumstances above referred to.”
He here quotes Sections 672 and 676 of Deady’s Code, being Sections 694, 698, B. &. C. Comp., and says:
“These provisions of the statute are declaratory of the law upon the subject, and are binding upon -the court. They limit the right of a party in the introduction of that character of testimony to those cases where the declaration forms part of the transaction which is in dispute, and provide that it is evidence as part of it. * * It occurs to me that courts at nisi prius would have but little difficulty in determining when the statements of a party in such cases were admissible as a part of the res gestae, or were incompetent upon the grounds that they were only hearsay, if they would consider whether the transaction to which they related was continuing when' they were made, or terminated at the time, and make that the test of the matter; and I believe that much of the embarrassment they labor under in applying the rule in such cases has arisen in consequence of an attempt that has frequently been made to stretch the res gestae doctrine to an unnatural extent in order to suit some supposed meritorious case, and which has led to the great diversity of decisions and confusion of the law upon that subject.”.
This decision has been cited by this court with approval on this question in Thomas v. Herrall & Zimmerman, 18 Or. 546 (23 Pac. 497), and Johnston v. O. S. L. Ry. Co. 23 Or. 94 (31 Pac. 283). An article upon the question of res gestae in 24 Cent. Law Journal, p. 463, quotes at length from the case of Sullivan v. O. R. & N. Co. supra, and from Waldele v. N. Y. C. & H. R. R. Co. 95 N. Y. 274 (47 Am. Rep. 41), and, referring to the Sullivan case, says:
“The test here suggested for the application of the doctrine of res gestae seems to be a very good and correct *40one, and which, I think, should command the considerate attention of every one having occasion to apply the doctrine. If followed, it would result in certainty and uniformity where there is now uncertainty and confusion.”
In concluding his article the author says:
“I confidently believe that the proper application of the rule or doctrine of res gestae, and some pertinent reasons for such application, will be found in the two cases (Sullivan v. O. R. & N. Co. and Waldele v. N. Y. C. & H. R. R. Co., 95 N. Y. 274: 47 Am. St. Rep. 41), above copiously quoted from, and that very much aid will be received in tne application oí uie doctrine in following their reasonings and suggestions.”
The case of L. & N. R. Co. v. Pearson, 97 Ala. 211 (12 South. 176) lays down practically the same rule as in the case of Sullivan v. O. R & N Co. 12 Or. 392, 398 (7 Pac. 508, 512: 53 Am. Rep. 364), and states- in conclusion:
“The real inquiry is: Did the main act proprio vigore further assert itself and demonstrate its character .or intent by impelling the contemporaneous or subsequent declaration or act, offered in evidence, and without which the main act is left incomplete and only partially proven, or did the declaration or circumstance offered as res gestae originate from some cause extraneous to the main act? If traceable solely to the main act, as the producing cause, and the declaration or circumstance as illustrative of the main act, it is res gestae; otherwise it is mere hearsay or irrelevant and inadmissible as res gestae.”
Waldele v. N. Y. C. & H. R. R. Co. 95 N. Y. 274 (47 Am. Rep. 41) contains a review of some of the cases on this subject, and says:
“The question is: Did the proposed declaration accompany the act, or was it so connected therewith as to constitute a part of it? If so, it is part of the res gestae, and competent; otherwise, not.”
Redman v. Metropolitan St. Ry. Co., 185 Mo. 1 (84 S. W. 26; 105 Am. St. Rep. 558) is to the same effect.
4. The statements of Smith were made to the witnesses Olsen and Mahoney, after he had left the engine and gone *41to the hold of the ship and was asked what was the matter. From his answers and statements it is evident that he feared he would be blamed for the accident, and sought to excuse himself. It was neither a spontaneous expression impelled by the act while operating the winch and characterizing it, nor was it such an expression of what took place in the hold of the ship as demonstrating its character; but it was a subsequent narrative or opinion made in response to an inquiry. Therefore the evidence of statements made by Smith, the engineer, was neither admissible as declarations of an agent to affect the principal nor were they res gestae; and, even if the evidence was sufficient to be submitted to the jury as to whether the dropping or jerking of the load in the manner complained of was occasioned by defective machinery—which we do not concede—still there was no evidence tending to show that the loads had so dropped or jerked, prior to the morning of the accident. Defendant himself says he did not know of it prior to the dropping of the load by which he was injured; and there is nothing in the evidence indicating that knowledge thereof was communicated to the defendants, or that the conditions were such that the defendants must be charged with notice of it: Kincaid v. O. S. L. Ry. Co. 22 Or. 35 (29 Pac. 3); Madden v. Occidental S. S. Co. 86 Cal. 445 (25 Pac. 5). The burden was upon plaintiff to show the negligence relied upon. This is not a case where proof of the accident prima fade shows negligence. There must be some affirmative proof of it: Woods, Master & Servant, § 382.
5. The plaintiff suggests, also, that, as the motion for the nonsuit was at first overruled, the court erred in sustaining it at the close of all the testimony, relying upon the case of Ferrera v. Parke & Lacey, 19 Or. 143 (23 Pac. 883). That .case, however, is not in point, for the reason that there the motion was not made until the evidence was all in and the ruling was based upon the result of all *42the evidence of both plaintiff and defendant, and this court held that the plaintiff had made a prima facie case, and, that being true, it became a question for the jury as to whether the defense was sufficient to overcome it; but in the case before us the trial court changed its ruling upon the admissibility of part of plaintiff’s evidence, namely, the statements of Smith, the engineer, in effect striking them out, and in sustaining the motion, held that upon plaintiff’s evidence alone, plaintiff had not made a prima facie case. Probably the better practice is to strike out the evidence erroneously admitted, and then direct a verdict; but the result of the nonsuit was more favorable to the plaintiff than a directed verdict would have been: Carroll v. Grande Ronde Electric Co. 49 Or. 477 (90 Pac. 903).
We find no errors in the ruling of the court below; and the judgment is affirmed. Affirmed.