------------------------+-----------+------------------

-----+
|           |    After heating to   |
Gum Arabic              |   Without |-------+-------+-------+
10 Per Cent.            |    heat.  | 60 deg.C. | 80 deg.C. | 100 deg.C |
------------------------+-----------+-------+-------+-------

+
Z at 18 deg.C               |    570    |  468  |  470  |  517  |
Z at 30 deg.C               |    485    |  400  |  422  |  439  |
Z at 50 deg.C               |    347    |  287  |  258  |  301  |
Ghatti gum No. 15,      |           |       |       |       |
5 per cent.  Z at 18 deg.C. |  1,104    |  780  |  660  |  758  |
------------------------+-----------+-------+-------+-------

+

The variation of viscosity with strength of solution was also studied with one or two typical gums.  A 10 per cent. is invariably more than twice as viscous as a 5 per cent. solution.  The following curve was obtained from one of the Ghattis.  Similar results were shown by other gums.

[Illustration:  Variation of Viscosity, with Dilution.  Ghatti No. 888.]

It would seem, therefore, that strong solutions, say of 50 per cent. strength, would be more alike in viscosity than solutions of 5 per cent. strength of the same gums.  In other words, the viscosity of a gum solution should be taken as nearly as possible to the strength it is used at, to obtain an exact quantitative idea of its gumming value.

The observation of this fact was one of the circumstances which decided us to use 5 per cent. solutions for the determination of Ghatti gum viscosities, the ratio between the 5 per cent. and 10 per cent. solutions of gum arabics being roughly the same as that between the respective weights required for gumming solutions of equal value.

From observation of the general nature of the solutions of Ghatti gums, and from the fact that when allowed to stand portions of the apparently insoluble matter passed into solution, the hypothesis suggested itself that metarabin was soluble in arabin, although insoluble in cold water.  If this hypothesis were correct, it would explain the apparent anomaly of Ghattis giving solutions of higher viscosity than gum arabics, although they leave insoluble matter behind.  The increase in viscosity would be due to the thickening of the arabic acid by the metarabin.  Moreover, the solutions yielded by various Ghattis leaving insoluble matter behind would be all of the same kind, viz., a saturated solution of metarabin in arabin more or less diluted by water.  Still further, if the insoluble residue of a Ghatti be the residual metarabin over and above that required to saturate the arabin, then it will be possible to dissolve this by the addition of more arabin in the form of ordinary gum arabic.  In order to see if this were the case the following experiments were performed.  Equal parts of a Ghatti and of a gum arabic were ground up together and dissolved in water.  The resulting solution was clear.  It was diluted until of 10 per cent. strength, and its viscosity then taken: