800 



THE AMERICAN NATURALIST [Vol. XL 



In the Arcella test the hexagons have no sides in common. Instead, 

 the hexagonal areas are so placed that the three adjacent sides of 

 three neighboring areas enclose a small triangular space. Just 

 here we find a further complication of the structure. These 

 interpolated triangles are not solid portions of the network, but 

 themselves contain areoles of subtriangular outline. The density 

 of the medium through which the light is transmitted seemed, 

 with the best illumination obtainable, the same in the small tri- 

 angular areoles as in the larger hexagonal areas. From this we 

 concluded that the areoles are dei)ressed areas in the network 



similar, excej)t ii> point of si/c and shape, to the hexagonal areas. 

 I)i:.i,n-an.in;iticall> thru, the nrtwork may l,e conceived as formed 

 of -^ftaiulit in t lu'cc >fi> ( .f p:i r;il!('ls, the lines of each set mak- 



ing an an-lr of ^i\t> d<-nv. uitl, ihu.c of the two other sets (see 

 diagram. Fig. .V coinpariM.n will, tlio actual pliotograplis, 



especially Fig. 2, seems to bear onl this conchision. That no aif 

 bubbles formed in the smaller spaces is natural, since i\w surround- 

 ing areas are much larger and of ecpial depth. 



