1. The percentage of water in the 14 dry varieties of glue examined varies between 9.0 and 21. The loss of water from isinglass is surprisingly large, especially as it cannot be explained by an artificial admixture of water, since the six varieties examined reabsorbed the same percentage of water from the air. The percentage of water in the other varieties of glue differs but little.
2. The various varieties of glue required different quantities of tannic acid for their precipitation, the amount for 100 parts of glue varying between 66 parts and 76.2, or calculated to per cent., between 49.4 and 56.8.
3. Placed in cold water, glue swells up and absorbs from 12 to 40 parts of water. The behavior of the various varieties differs very much in this respect, and in most of the experiments, with the exception of variety No. 4, the percentage of glutin is in exact proportion to the quantity of water absorbed.
4. The strength of the gelatinized glue varies between 12 grammes (185.18 grains) and 64 grammes (987.67 grains) for a 10 per cent. solution. This property corresponds neither with the absorption of water nor the percentage of glutin.
Variety No. 4, in the table, contains 52 per cent. of glutin, but does not gelatinize, its strength being therefore equal to 0, while variety No. 14, with 49.4 per cent. of glutin, therefore less than No. 4, shows a strength of 42. As no close connection between the properties mentioned in the table can be recognized, it is best not to be guided by these properties alone, but to test also the behavior of the glue in practical use. Such test consists in estimating the adhesive power of the glue from the weight required to tear asunder two pieces of wood glued together and dried. But as the results obtained by this purely practical test must necessarily vary on account of the impossibility of having two surfaces of wood always exactly alike, and the uncertainty of applying every time the same quantity of glue, Weidenbusch’s method may here be given. It is based upon the principle that sticks of plaster of Paris cast of exactly the same material and of uniform size break under the same weight when placed upon two supports and loaded in the centre. If now such plaster sticks be saturated with glue solutions prepared according to the same method, but from different qualities of glue, a greater force will be required for breaking them, and this force will be the greater the better the quality of glue is.