880 



THE AMERICAN NATURALIST 



[Vol. XL 



cytoplasm, occasionally projects into the cavity which penetrates 

 the branch. The gill cavities occupy the main filaments, pene- 

 trate the branches, and communicate freely with the body cavity 

 at the base of the gills. Found within these chambers are masses 

 of slender trachese and as their presence in these organs strongly 

 suggests that they do not function wholly as blood gills, I shall dis- 

 cuss them in detail. 



During the larger part of the time I spent in this work, I believed 

 tracheae to be wanting in these gills, but masses of fine tube-like 

 structures which appeared in the sections excited doubt. Having 

 secured some living larvae, I cut off, after many unsuccessful 

 attempts, a projected gill. This being quickly mounted and exam- 

 ined with a l^-inch objective and 1-inch ocular, there appeared 

 reaching up into the gill cavities, masses of fine tracheal tubes 

 filled with air. While they enter all parts of the basal opening, 

 they are thickest in four main bands, two entering the central lobe 

 and one each of the lateral lobes. This collection into bands 

 might result from mechanical causes but, as a similar arrangement 

 appeared in the sectioned material, it seems probable that it is the 

 natural condition. The tubes are small, about 1 fi in diameter, of 

 uniform size, and rarely branching unless in contact with the wall. 

 They penetrate all parts of the gill cavity, and generally one and 

 frequently two slender tracheae lying close to the wall penetrate to 

 the very tip of a unicellular branch. When viewed with a A oil- 

 immersion lens there may be seen in mounted and stained sections, 

 given off from these tracheae which lie close to the wall, fine branches 

 penetrating the protoplasm. These air vessels gradually decrease 

 in size as they approach the ends of the cavities. Just below the 

 base of the gills they average 2,m in diameter, half way up the main 

 lobe they are about 1 /i, anrl near the ends of the finger-like branches 

 they have decreased to a diameter of O..S/(. These measurements 

 were taken while the mhcs \ven> iWlvd with air. 



The trachea' branch very rarely in tlie o-ill cavity and in such 



