208 ANNUAL REPOET SMITHSONIAN INSTITUTION, 1917. 



Many different kinds of food were offered corals, but they took only 

 animal food — they are entirely carnivorous. The following experi- 

 ment was tried many times : A piece of diatom mat was placed on one 

 side of the oral disk and a piece of crab meat on the other. In- 

 variably the crab meat was seized and swallowed ; while the diatoms 

 induced no reaction except ultimately to be removed from the sur- 

 face. No kind of purely vegetable food was taken by any one of the 

 numerous species investigated. However, pieces of plants coated 

 with small animals or soaked in meat juice will be swallowed, and 

 later the vegetal matter ejected. 



One of the experiments that I found particularly fascinating was 

 to drop living specimens of the small Crustacea, known as copepods, 

 within the expanded tentacles of Masandra areolata. These little 

 animals because of the quickness of their motion are popularly 

 called water fleas, but they did not move swiftly enough to escape 

 the lightning-like dart of the coral's tentacles and nematocysts — they 

 were all caught and swallowed. 



Plate 18 illustrates Maeandra areolata during the swallowing and 

 digestion of rather large pieces of food. The usual behavior of 

 corals while they are gorged with food and after hunger is entirely 

 satisfied, is to retract their tentacles and other distensible organs. 

 Furthermore, after complete satiation, the direction of ciliary mo- 

 tion reverses and particles of food dropped on the surface will be 

 moved toward the periphery of the coral in a manner similar to that 

 in which the surface is cleaned of nonnutrient particles. 



The distribution of corals according to depth is dependent upon 

 the distribution of small floating and swimming animals which 

 entirely supph" their food. Should the quantity of such food de- 

 crease wdth increasing depth, such decrease would limit the down- 

 ward extent of the shoal-water fauna, but as I do not know of any 

 quantitative estimates of the amount of such food above and below 

 46 meters in coral reef areas, there is no basis for a positive opinion. 



REARING CORAL LARVAE. 



The rearing of coral larvae is important because only by knowing 

 the duration the free-swimming planulae stage can the possibilities 

 of the distribution of corals by marine currents be understood ; and, 

 it is obvious that, in order to ascertain precisely how rapidly corals 

 grow, the life of the same colony should be followed from the time 

 the planula to w^hich it owes its existence first settled. Two of the 

 methods of obtaining and rearing planulae will be briefly described. 



Colonies extruding planulae were brought into the laboratory and 

 kept in glass vessels of sufficient size to furnish an adequate supply 

 of water, which should be changed rather frequently. The planulae 

 were removed from the vessels containing the parent colony to a 



