124 ON SOME IRIDESCENT FILMS [36 plates was unknown this conjecture could not be satisfactorily confirmed No ordinary cleaning or wiping had any effect; to remove the films recours must be had to hydrofluoric acid, or to a polishing operation. My frien< Prof. T. W. Richards, after treating one with strong acids and other chemicals pronounced it to be what chemists would call " very insoluble." The plate first encountered manifested (in the air) a brilliant glassy surface, bu afterwards I found others showing in the water nearly or quite as goo< colours, but in the air presenting a smoky appearance. Desirous of obtaining the colours as perfectly as possible, I endeavoure< to destroy the reflexion from the back surface of the plate, which would I supposed, dilute the colours due to the iridescent film. But a coating o black sealing-wax, or marine glue, did not do so much good as had bee] expected. The most efficient procedure was to grind the back of the plate as is very easily done with carborundum. The colours seemed now to be a good as such colours can ever be, the black also being well developed. Doubtles the success was due in great measure to the special localized character of th illumination. The substitution of strong brine for water made no perceptibL improvement. At this stage I found a difficulty in understanding fully the behaviour o the nnground plates. In some places the black would occasionally be good while in others it had a washed-out appearance, a difference not easib accounted for. A difficulty had already been experienced in deciding upoj which side of a plate the film was, and had been attributed to the extremi thinness of the plates. But a suspicion now arose that there were films upoi both sides, and this was soon confirmed. The best proof was afforded b^ grinding away half the area upon one side of the plate and the other half o the area upon the other side. Whichever face was uppermost, the ungroum half witnessed the presence of a film by brilliant coloration. Attempts to produce silicate films on new glass were for some time ai almost complete failure. I used the formula given by Abney (Instruction ii Photography, llth edition, p. 342):— Albumen ................................................ 1 part. Water ................................................... 20 parts. Silicate of Soda solution of syrupy consistency 1 part. But whether the plates (coated upon one side) were allowed to drain and drj in the cold, or were more quickly dried off over a spirit flame or before a fire the resulting films washed away under the tap with the slightest friction o: even with no friction at all. Occasionally, however, more adherent patchei were observed, which could not so easily be cleaned off. Although it did no seem probable that the photographic film proper played any part, I triec without success a superposed coat of gelatine. In view of these failure,lass rendering it optically different from the interior.