PER CURIAM.

We assume arguendo that both the trade-marks, "Bouquet des Amours" and "Fleurs d'Amours," are proper marks of origin, acquired by valid user, and that the Trade-Mark Convention of April 19, 1869, did not invalidate them. Granting so much, in our judgment the plaintiff has failed to prove any infringement of either by the use of "Charme d'Amour," or of "Caresse d'Amour." Prima facie, the source of origin of the products sold under them is indicated by the mark taken as a whole; one who uses less than that whole may perhaps infringe, but, if so, it must appear that the part he has taken identifies the owner's product without the rest. There are, indeed, cases where the change is so slight that judges will infer as much by mere inspection, though even there the question is always one of fact. In the case at bar we have no antecedent reason to assume that the word "Amour" alone meant Roger Gallet; it had been often applied to perfumes before, and its appropriateness is plain at a glance.

Thus the plaintiff had to prove the issue independently of the word common to all the marks. An appellate court of equal authority upon nearly the same showing has decided against it; we will not lightly ignore that holding. Besides, the evidence greatly preponderates against the conclusion, the defendants having gone further and amply proved that in fact no confusion existed. We shall not discuss this evidence in detail; the several earlier tribunals who have dealt with the issue have considered it, and we are satisfied with their conclusions. All we need mention is the new evidence, consisting of the orders received by the plaintiff from its two customers, which were excluded below on the ground that they were self-serving declarations. That, of course, they were not, and, moreover, they seem to us, all things considered, to have been enough authenticated to be competent. The plaintiff is entitled, therefore, to insist upon their consideration, especially as they are the only evidence of confusion worth notice.

On their face they indeed appear to be inquiries from customers, and tradesmen at that, who were misled into supposing that the defendants' goods came from the plaintiff. While the defendants would, it is true, be entitled to present their side of the matter, which the exclusion of the proof prevented, we must take them at their face for the purposes of these appeals. It is to be observed that they are not instances of persons seeking the plaintiff's product and getting the defendants', the only wrong of which it can complain. Their cogency arises only from the inference that, if tradesmen supposed that the plaintiff made "Amour Encharm" (sic) and "Caress d'Amour" (sic), a fortiori the public must do so too, and would buy them under that misapprehension. That is, indeed, not an unreasonable inference; that is to say, the inquiries were legitimate evidence on the plaintiff's side. But for all that they were not necessarily controlling. Taken with the proof as a whole, especially with Jordan's admission that he had never known of any confusion, they do not seem to us enough to turn the scale. Some buyers will be confused by the most remote similarities, and, while the owner has an equal interest in his careless customers, there is practically a limit to which in the protection of that interest he can control the use of words even of foreign origin. It is a commonplace, especially in this branch of the law, that rights emerge only upon a balance of conflicting interests. Registry changes nothing, and is important only for procedural purposes. The inference of substantial confusion from these two inquiries seems to us overborne by the defendants' evidence, and a decree for the plaintiff would have to rest upon the unproved assumption that the plaintiff was in fact suffering any perceptible loss of trade because of its customers' mistakes.

Decrees affirmed.