Another danger lies in the instability of the pollen extracts. Koessler expressly warns against commercial preparations of pollen protein because of the danger of decomposition. His extracts do not keep more than three weeks and are dangerous to use after that time. On the other hand, Goodale, making his extracts with 15% alcohol, reports them as active and fit for use after more than one year. Oppenheimer and Gottlieb object to commercial preparations on different grounds. The commercial preparations contain many different pollens so as to be sure to include those to which the patient is sensitive. They point out that in these mixtures, the dose of the individual pollens cannot be adjusted to the changing conditions of the patient and, in addition, injecting into the blood of the patient pollens to which he is not already sensitive may sensitize him to these pollens also and leave him worse than before. There is the lesser danger that the patient will not be immunized by the injections but become more sensitive to his old pollens than he was before, as Noon pointed out in his first paper; for artificial immunity is a difficult thing to control and is by no means as easy as it looks in the book. Nor is it as easy to immunize a human being over many years of life, subject to so many conflicting influences, as it is to immunize a guinea-pig living in a cage. To lessen the dangers and enable the patient to keep up his immunization for many months and years, Goodale borrowed an idea from Schloss, who fed his egg-oat-meal-almond anaphylaxis patient minute doses of these foods until he so raised his resistance that he was able to eat them in ordinary quantities without harm. Such artificial resistance must be kept up by eating a small quantity of the food each day or it will be lost (old King Mithridates again). As most of the pollens are not edible and as patients react to botanically allied plants, Goodale tried feeding the patient over long periods of time with vegetables and edible plants that were allied to the offending pollens, expecting that, as in Schloss's patient, some minute part of the protein would pass unto the blood unchanged and maintain the protection. So far, these feeding experiments have failed. Trial with homoeopathically potentized pollens over long periods of time has not been made.