The purpose of this work was to compare the respiration rate of the leaves of the healthy paulownia tree and that of the diseased one infected with witche's-broom, and to discuss what really caused the death of the diseased tree. It was confirmed from the results of the former experipents by the same author that the respiration rate of all the leaves above the fourth leaf order (from the top) was similar to that of all the leaves of the twig. Therefore, in this experiment, all the leaves above the fourth leaf order (from the top) were used as an experimental materials in order to compare the respiration rate of the diseased tree with that of the healthy one. The respiration rate was determined by the Warburg macro-manometric method. The experiments were carried out from July to November. The monthly measurement with the diseased and healthy trees were taken at the same time and nine times respectively. (1) Qo_2 (the quantity of oxygen consumption) of the diseased leaves, in July, is about 70 percent greater than that of the healthy one. Highly significant difference is recognized statistically in this case. (2) Qo_2 of the diseased leaves, in August, is about 20 percent greater than that of the healthy one. Qco_2 (the quantity of carbon dioxide production) of the diseased leaves is about 36 percent greater than that of the healthy one. RQ (ratio of Qco_2 to Qo_2) of the diseased leaves is about 13 percent greater than that of the healthy one. Highly significant differences are recognized statistically in these three cases mentioned above. (3) Qo_2 of the diseased leaves, in November, is about 29 percent greater than that of the healthy one. Qco_2 of the diseased leaves is about 40 percent greater than that of the healthy one. In these cases highly significant differences are recognized statistically. RQ of the diseased leaves about 8percent greater than that of the healthy one in this case significant difference is recognized statistically. (4) The respiration rate of the leaves injured by a knife increases approximately twice as much as that of the non-injured ones immediately following the injury, and it gradually decreases with the lapse of time. It is necessary to let the injured leaves stand at least three hours or more before they are used as the experimental materials for respiration studies. (5) If the respiration is considered as the consumption of energy, the results of the experiments indicate the greater consumption of energy in the diseased leaves than that the healthy ones. Cosequently, it is concluded that the diseased tree will come to die, when the consumption of energy (respiration) of the tree overcome the accumulation of energy (photosynthesis) in one physiological macro-unit (a twig). The differences in RQ between the diseased leaves and healthy ones seems to indicate the abnormal change taking place in the metabolism of the diseased ones.