I may add that the stomach itself appears to do very little, possibly nothing, towards the digestion of starch. The primary conversion into dextrin is effected by the saliva, and the subsequent digestion of this takes place in the duodenum and following portions of the intestinal canal. This applies equally to the less easily digested material of the vegetable tissue described in the preceding chapter. Hence the greater length of the intestinal canal in herbivorous animals as compared with the carnivora. Having described the changes effected by heat upon starch, and referred to its further conversion into dextrin and sugar, I will now take some practical examples of the cookery of starch foods, beginning with those which are composed of pure, or nearly pure, starch. When arrowroot is merely stirred in cold water, it sinks to the bottom undissolved and unaltered. When cooked in the usual manner to form the well-known mucilaginous or jelly-like food, the change is a simple case of the swelling and breaking up of the granules already described as occurring in water at the temperature of 140° Fahr. There appears to be no reason for limiting the temperature, as the same action takes place from 140° upwards to the boiling point of water. I may here mention a peculiarity of another form of nearly pure starch food, viz. tapioca, which is obtained by pulping and washing out the starch granules of the root of the _Manihot_, then heating the washed starch in pans, and stirring it while hot with iron or wooden paddles. This cooks and breaks up the granules, and agglutinates the starch into nodules which, as Mr. James Collins explains (‘Journal of Society of Arts,’ March 14, 1884), are thereby coated with dextrin, to which gummy coating some of the peculiarities of tapioca pudding are attributable. It is a curious fact that this _Manihot_ root, from which our harmless tapioca is obtained, is terribly poisonous. The plant is one of the large family of nauseous spurgeworts (_Euphorbiaceæ_). The poison resides in the milky juice surrounding the starch granules, but being both soluble in water and volatile, most of it is washed away in separating the starch granules, and any that remains after washing is driven off by the heating and stirring, which has to reach 240° in order to effect the changes above described. I suspect that the difference between the forms of tapioca and arrowroot has arisen from the necessity of thus driving off the last traces of the poison, with which the aboriginal manufacturers are so well acquainted as to combine the industry of poisoning their arrows with that of extracting the starch-food from the same root. No certificate from the public analyst is demanded to establish the absence of the poison from any given sample of tapioca, as the juice of the Manihot root, like that of other spurges, is unmistakably acrid and nauseous.