1902.] NATURAL SCIENCES OF THILADELPHIA. 287 



renewal of water in the filter- paper, aud care was takeu to leave uo 

 large drops of water iu the bottom of the dish.* 



Several observers have remarked the prolonged starvation which 

 the isopods can withstand (Ide (3), Conklin (4), McMurrich (5), 

 Schonicheu (6) ). Xo attempt has been made here to determine the 

 ultra- maximum time; but in the course of the study an isolated ani- 

 mal has frequently been starved for more than a month. When 

 several are starved together in the same dish, the time of absolute 

 deprivation seldom reaches a month ; some are sure to moult, or to 

 become greatly weakened, or to die from some cause, when the legs 

 and soft parts are then eaten by the others. While this does not 

 necessarily terminate the fast for all, it does make impossible that 

 determination of the fasting period which is often desirable. The 

 only way to obtain this datum with certainty is to isolate the animals. 

 For this purpose small flat-bottomed vessels of various dimensions, 

 preferably all of the same height, so that they could be covered with 

 the same sheet of glass and filter-paper, have been employed. 



Various methods of removing the intestine have been tried. 



*The latter precaution was necessary to guard against drowniug.^ 

 When by chance the breathing appendages are covered with water, it 

 the animal is m its natural habitat or on filter-paper or other object 

 which will take up the water readily, it very rapidly frees Uselt by 

 dipping the abdomen, just as one might remove an extra drop ot mk 

 from a fountain pen by touching it to an absorbing surface. Kept on 

 clean glass, however, the drop is not always removed, and the animal 

 may die from suffocation or over-exertion, or both. 



A greater difficulty was experienced iu cases of prolonged starvation 

 from cannibalism. Although iu the main vegetable feeders, or scavengers 

 at most, and particularly iuofteusive toward other species of animals, 

 these Crustacea will eat one another, if driven by hunger to this extremity. 

 While some signs of offensive attack have been seen, such as a vigorous 

 nip by one animal at the dorsum of another, plainly sufficient to inflict 

 considerable pain, it would be unsafe to regard an act of this kind as a 

 deliberate attempt to kill. Moreover, very young animals have been 

 kept alive in the same dishes with adults through a month or six weeks 

 of starvation. Individuals isolated in small stender dishes have often 

 been seen to moult and to survive the ordeal just as in the state of 

 nature. But if one moult (Sec. iii, (4) ) in a dish with other starving ani- 

 mals or is greatly weakened by its efforts to remove a drop of water, it 

 may fall an easy victim to the hunger of its starving companions. Ihe 

 cannibalism to be guarded against is, tlierefore, of an accidental sort, 

 and arises from an instinct to keep the intestine fall. The same instinct 

 prompts the animal in similar circumstances to eat anything it_ can 

 swallow, although it is not intended by this to imply that no choice is 

 exercised in the selection of its natural food. 3Iention will be made in a 

 later section of the care necessary to induce animals which have been 

 fasting a long time to eat the pure foods with which it is de>irable to feed 

 them. 



