I concur in the conclusion reached by the majority of the court, but I do not find in the Fitch & St. John patent, nor in any other patent or publication prior to March, 1892, a sufficient disclosure of the principle of Gunn’s device to negative patentable novelty. So far, as all prior literature is concerned, ft seems to.me that he was the first to devise a combination which would. *437enable one to arrange the cards under a double classification within the same space. An illustration will best show the change he accomplished. Suppose it were desired to classify lor reference by means of a card index the various judicial decisions in admiralty, one decision on each card. The various divisions of the subject — e. g., “charter party,” “insurance,” “navigation,” etc. — would be indicated by division cards each with projecting tab appropriately lettered, each of such cards being followed by the recording cards giving decision and citation of the authority. Tor subdivision under “navigation” other division cards with the tab differently placed would be similarly used for “collision,” “towage,” etc. And in like manner “collision” could be divided into “ferry boats,” “steam vessels,” “sailing vessels,” etc. Such a single method of division and subdivision might be continued as far as required; but, if it were desired to have all the decisions so classified that at the same time they were divided alphabetically — e. g., “Andrews v. Smith,” “Brown v. S. S. Triton,” etc. — none of the card indexes of the prior patents would admit of such a double classification, and no prior patent or publication, as it seems to me, would indicate to one skilled in the art how it might be accomplished. By Gunn’s system it is easily done. The recording cards may be arranged in alphabetical order of the plaintiff’s names (with a division card for each letter of the alphabet), and the other divisions and subdivisions indicated by tabs on the recording cards themselves appropriately positioned. I should be inclined therefore to find patentable novelty were it not for the prior use known as the Curtis Publishing Company exhibit. That places classification marks on the recording cards themselves, while at the same time division cards may be used for effecting double classification. The distinguishing mark in Curtis exhibit is color, instead of differently positioned tabs, and it is not so efficient as Gunn’s device, because it is more difficult to pick out colors and the range of differences is perhaps less, but the exhibit does show the fundamental change which differentiated Gunn’s improvement broadly from the earlier art, and with the exhibit before him it did not require invention for one skilled in that art to elaborate further improvements calculated to make the card index more convenient for use.