THE RELATION OF DIET TO DISEASE 425 
majority among the rarer specimens, ten carnivorous, 
seven herbivorous and one omnivorous mammals, ten 
carnivorous and two seed-eating birds, due possibly to 
inappropriate diet or to some unknown factor that ren- 
dered the diet inadequate. At the autopsy nothing was 
found to account for death except the draining of all 
storage supplies. 
Staevation". 
The reports of studies conducted during long labora- 
tory fasts have been among the most valuable records for 
the understanding of the chemical requirements of diet 
and of the close chemical interrelationship existing 
between the different food factors. In absolute starva- 
tion life is very short, primarily because water is 
necessary for respiration, for dissolving products of 
metabolism and for preventing changes in digestive intes- 
tinal secretions. The amount of water needed varies with 
different species of animals. If the water is supplied, the 
organism is enabled to maintain its energy for continued 
existence from the destruction of its own tissues. The 
length of life depends upon the amount of protein 
ingested before the fast commenced, and the amount of 
stored fat and glycogen, especially that stored in the liver. 
The mechanism of the results is similar. The animal 
body uses first its available glucose, and when this is 
partially exhausted bums its stored fat and protein. 
The fat combustion is usually defective, ketone bodies 
appearing in the urine in large quantities. The change 
from fat to protein metabolism accounts for the pre- 
mortal rise in metabolism which occurs usually a few days 
before death. The chemical composition and corpuscular 
richness of the blood is tenaciously preserved; glucose 
and protein concentration are practically normal up to 
the day of death. There is at times a slight increase in 
globulins and always an increase in fat due to its trans- 
portation from storage depots. The cause of death is 
28 
