As cooling may result in increased sugar in the blood, and consequent glycosuria, the rectal temperature was observed from time to time, and it was found to vary so slightly that in these experiments it was a wholly negligible factor. In one cat the rectal temperature fell to 36° C. while the animal was bound and placed in a cold room (about 2° C.) for fifty minutes, but no sugar appeared in the urine. Further evidence that the appearance of sugar in the urine may arise purely from emotional excitement was obtained from three cats which gave negative results when bound in the holder for varying periods up to four hours. It was noteworthy that these animals remained calm and passive in their confinement. When, however, they were placed, separately, in a small wire cage, and were barked at by an energetic little dog, that jumped at them and made signs of attack, the cats became much excited, they showed their teeth, humped their backs, and growled defiance. This sham fight was permitted to continue for a half hour in each of the three cases. In each case the animal, which after four hours of bondage had exhibited no glycosuria, now had sugar in the urine. Pain, cooling, and bondage were not factors in these experiments. The animal was either frightened or enraged by the barking dog, and that excitement was attended by glycosuria. The sugar excreted in the twenty-four hours which included the period of excitement was determined by the Bertrand method.[60] It ranged from 0.024 gram to 1.93 grams, or from 0.008 gram to 0.62 gram per kilo body weight, for the twenty-four hours’ quantity. The presence of sugar in the urine may be used as an indication of increased sugar in the blood, for unless injury has been done to the cells of the kidneys, they do not permit sugar to escape until the percentage in the blood has risen to a considerable degree. Thus, though testing the urine reveals the instances of a high content of blood sugar, it does not show the fine variations that appear when the blood itself is examined. Recently Scott[61] has concluded a thorough investigation of the variations of blood sugar in cats, and has found that merely incidental conditions, producing even mild excitement, as indicated by crying or otherwise, result in a noticeable rise in the amount. Indeed, so sensitive is the sugar-liberating mechanism that all the early determinations of the “normal” content of sugar in blood which has been drawn from an artery or vein in the absence of anesthesia, are of very doubtful value. Certainly when care is taken to obtain blood suddenly from a tranquil animal, the percentage (0.069, Scott; 0.088, Pavy) is much less than when the blood is drawn without anesthesia (0.15, Böhm and Hoffmann), or after light narcosis (0.282, Rona and Takahashi[62]).