4. The proportion of iron in the precipitate varies greatly with the length of time the ink has been exposed. At first the precipitate contains 10 per cent of iron, but by and by a new one having only 7.5 per cent is formed, and in from forty to seventy days we find one of 5.7 per cent. Simultaneously iron increases in the ink (proportionate to the tannin). 5. The results show, and practice confirms, that 16 parts of iron (80 ferrous sulphate) and 100 parts of tannin are best for ink manufacture. The research now travelled in a direction which accumulating experience showed to be obligatory. Blue-black tannin ink lost color, and the reducing nature of the tannin tended to the formation of a highly objectionable precipitate in the ink, which made writing anything but a pleasure. These two faults were doubtless linked together in some way and seemed not to exist when gallic acid was used, for ink so made was found to precipitate only after a long exposure, it required no free acid to keep the precipitate in solution, and retained the indigo blue color for a long time; alkalis did not decompose the ink, and provided blacker and more permanent writing. Determination of the correct proportions of gallic acid and ferrous-sulphate was the subject of prolonged experiments conducted on similar lines to those already detailed. The conclusions as to precipitation were also similar. Thirty parts of iron (150 of ferrous- sulphate) and 100 parts of gallic acid were found to be the most suitable proportions for ink-making. It is advisable, however, not to discard tannin altogether, owing to the slow blackening of the gallic acid ink, and a little tannin gives initial blackening and body, while it is absolutely necessary for copying ink. Initial blackness can also be ensured by oxidizing 21 per cent of the ferrous-sulphate without adding the extra acid necessary to the formation of a ferric salt. The concluding portion of his research is devoted to the influence of sugar upon the permanence of ink, and the results of the experiments are summed up in the following sentences: "It would be injurious to add 3 per cent of sugar to a tan in ink, while from 4 to 10 per cent would be quite allowable. Most copying inks contain about 3.5 per cent of sugar-- not far from the critical amount. With gallic acid more than 3 per cent of sugar hardly varies the precipitate, but the importance of this point is somewhat diminished by the fact that the presence of sugar is by no means necessary in a writing ink. Dextrin is a much superior substance to use. Curiously this body rapidly precipitates a tannin ink; hence it is useless for copying ink, but for the gallic ink it is an excellent thickener." Chen-Ki-Souen, "Lencre de China," by Maurice Jametel, appeared in Paris in 1882, but as the title indicates, it is the old "Indian" or Chinese ink that is discussed.